feat(gateway): external drain trigger + accept-gating (begin/cancel + control channel)

Tasks 2.1 + 2.2 + 2.3 of the safe-shutdown plan — the reversible
quiesce-without-restart machinery NAS drives during a lifecycle action (D4a).
These ship together because the endpoint, the control channel, and the gateway
state machine are one coherent slice.

2.2 — control channel (gateway/drain_control.py, new):
The dashboard has no HTTP path into a running gateway (guardrails: "there is NO
external control channel into a running gateway"); restart/drain is driven only
by markers the gateway reacts to. So begin/cancel-drain writes/removes a
presence-based marker .drain_request.json (HERMES_HOME-scoped, atomic write,
never-raises read; a corrupt marker reads as present-contentless → fail-safe
toward quiescing). This is Q-B option A.

2.2 — gateway state machine (gateway/run.py):
- _external_drain_active flag, DISTINCT from the shutdown _draining flag: this
  one does NOT exit the process and is fully reversible.
- _enter_external_drain / _exit_external_drain: idempotent transitions that
  flip gateway_state→draining / →running via _update_runtime_status (preserving
  the live active_agents count). exit refuses to revert to running during a
  real shutdown or after the loop stops (shutdown wins).
- _drain_control_watcher: 1s background task (modelled on _handoff_watcher)
  reconciling accept-state with the marker; honours a marker that survived a
  restart on its first tick. Registered alongside the other watchers in start.
- New-turn accept gate in _handle_message, placed BEFORE the session-slot
  claim: when draining, refuse to START a new turn (so active_agents can only
  fall → no TOCTOU race), while in-flight turns finish untouched. Internal/
  system events (restart-recovery replays, bg-process completions) bypass it.

2.1 — endpoint (hermes_cli/web_server.py):
POST /api/gateway/drain {action: drain|cancel}. Authenticated by the Task-2.0a
token seam (the drain plugin registered this exact path as a token route);
attributes the request to the verified token principal. Begin writes the
marker, cancel removes it — the gateway process owns the actual transition.
Force-override (D6) is NOT here; it maps onto the existing immediate
/api/gateway/restart force path.

Tests (mocked — necessary-not-sufficient; the HARD live gate Q-B is next):
- tests/gateway/test_external_drain_control.py — marker contract (write/clear/
  read/corrupt/atomic), state machine (enter/exit/idempotency/shutdown-wins/
  loop-stopped), watcher reconcile-enter-then-exit, new-turn refusal, and
  in-flight-not-interrupted. 15 tests.
- tests/hermes_cli/test_web_server.py — /api/gateway/drain begin/default-begin/
  cancel/cancel-idempotent/bad-action-400. 6 tests.
- dashboard.drain_auth config section already added in 2.0b commit.

All touched suites green: 301 (gateway+auth) + 9 (web_server endpoints) passed.

Intentionally deferred:
- HARD live-validation gate (Q-B): real isolated `hermes gateway run`, drive a
  real begin-drain marker, prove the 5-point checklist a–e.
- Spec-doc status flip + Phase-2 PR.

Build status: external-drain, restart-drain, status, dashboard-auth, drain-plugin,
token-auth, and web_server-endpoint suites green.

.dockerignore

@@ -66,12 +66,8 @@ runtime/
 
 # ---------- Not needed inside the Docker image ----------
 
-# Desktop app source (Tauri/Electron); never installed in the container.
-# apps/shared is the dashboard↔desktop websocket helper and is linked from
-# web/package.json as a file: workspace dep — keep it in the build context.
+# Desktop app source (Tauri/Electron); never installed in the container
 apps/
-!apps/shared/
-!apps/shared/**
 
 # Test suite — not shipped in production images
 tests/
@@ -97,6 +93,9 @@ packaging/
 plans/
 .plans/
 
+# ACP registry manifest (icon + agent.json) — not consumed at runtime
+acp_registry/
+
 # Repo-level dotfiles that are git-only or dev-tooling config
 .env.example
 .envrc

.env.example

@@ -1,13 +1,6 @@
 # Hermes Agent Environment Configuration
 # Copy this file to .env and fill in your API keys
 
-# =============================================================================
-# LLM PROVIDER (Fireworks AI)
-# =============================================================================
-# Get your key at: https://app.fireworks.ai/settings/users/api-keys
-# Address models directly by catalog ID, e.g.
-#   accounts/fireworks/models/kimi-k2p6, accounts/fireworks/models/glm-5p2
-# FIREWORKS_API_KEY=
 # =============================================================================
 # LLM PROVIDER (OpenRouter)
 # =============================================================================
@@ -115,10 +108,6 @@
 # HF_BASE_URL=https://router.huggingface.co/v1  # Override default base URL
 # OPENCODE_GO_BASE_URL=https://opencode.ai/zen/go/v1  # Override default base URL
 
-# DeepInfra — 100+ top open models, pay-per-use.
-# Get your key at: https://deepinfra.com/dash/api_keys
-# DEEPINFRA_API_KEY=
-
 # =============================================================================
 # LLM PROVIDER (Qwen OAuth)
 # =============================================================================
@@ -136,15 +125,6 @@
 # Optional base URL override:
 # XIAOMI_BASE_URL=https://api.xiaomimimo.com/v1
 
-# =============================================================================
-# LLM PROVIDER (Upstage Solar)
-# =============================================================================
-# Upstage provides access to Upstage Solar models.
-# Get your key at: https://console.upstage.ai/api-keys
-# UPSTAGE_API_KEY=your_key_here
-# Optional base URL override:
-# UPSTAGE_BASE_URL=https://api.upstage.ai/v1
-
 # =============================================================================
 # TOOL API KEYS
 # =============================================================================

.envrc

@@ -1,5 +1,5 @@
-watch_file pyproject.toml uv.lock hermes
+watch_file pyproject.toml uv.lock
 watch_file package-lock.json package.json web/package.json ui-tui/package.json website/package.json apps/shared/package.json apps/desktop/package.json ui-tui/packages/hermes-ink/package.json
-watch_file flake.nix flake.lock nix/devShell.nix nix/tui.nix nix/package.nix nix/python.nix nix/hermes-agent.nix nix/desktop.nix
+watch_file flake.nix flake.lock nix/devShell.nix nix/tui.nix nix/package.nix nix/python.nix
 
 use flake

.github/actions/detect-changes/action.yml

@@ -1,105 +0,0 @@
-name: Detect affected areas
-description: >-
-  Classify a PR's changed files into CI work lanes (python, frontend, site,
-  scan, deps, mcp_catalog) so the orchestrator can conditionally call only
-  the sub-workflows a PR can affect. Outputs are always "true" on push/dispatch
-  events and fail open (everything "true") when the diff cannot be computed.
-
-inputs:
-  github-token:
-    description: Token for the GitHub API (gh CLI). Pass steps.app-token.outputs.token from the calling workflow.
-    required: false
-    default: ${{ github.token }}
-
-outputs:
-  python:
-    description: Run Python tests / ruff / ty / windows-footguns.
-    value: ${{ steps.classify.outputs.python }}
-  frontend:
-    description: Run the TypeScript testing matrix + desktop build.
-    value: ${{ steps.classify.outputs.frontend }}
-  docker_meta:
-    description: Docker setup and meta files have changed.
-    value: ${{ steps.classify.outputs.docker_meta }}
-  site:
-    description: Build the Docusaurus docs site.
-    value: ${{ steps.classify.outputs.site }}
-  scan:
-    description: Run the supply-chain critical-pattern scanner.
-    value: ${{ steps.classify.outputs.scan }}
-  deps:
-    description: Check pyproject.toml dependency upper bounds.
-    value: ${{ steps.classify.outputs.deps }}
-  npm_lock:
-    description: Post/update the semantic package-lock.json diff PR comment.
-    value: ${{ steps.classify.outputs.npm_lock }}
-  mcp_catalog:
-    description: Require MCP catalog security review label.
-    value: ${{ steps.classify.outputs.mcp_catalog }}
-  ci_review:
-    description: Require CI-sensitive file review label.
-    value: ${{ steps.classify.outputs.ci_review }}
-  ci_review_files:
-    description: JSON list of CI-sensitive files changed by the pull request.
-    value: ${{ steps.classify.outputs.ci_review_files }}
-
-runs:
-  using: composite
-  steps:
-    - name: Classify changed files
-      id: classify
-      shell: bash
-      env:
-        # Fall back to the built-in read-only token when the caller passes an
-        # empty value. Fork PRs get no repo secrets, so AUTOFIX_BOT_PAT is ""
-        # there, and an input `default:` only applies when the input is omitted,
-        # not when it's passed empty. Without this fallback the compare API
-        # fails on forks and the classifier fails open (every lane forced on).
-        GH_TOKEN: ${{ inputs.github-token || github.token }}
-        REPO: ${{ github.repository }}
-        EVENT_NAME: ${{ github.event_name }}
-        BASE_SHA: ${{ github.event.pull_request.base.sha }}
-        HEAD_SHA: ${{ github.event.pull_request.head.sha }}
-      run: |
-        set -euo pipefail
-
-        # Only pull_request events are gated. Other events (push, release,
-        # dispatch) leave CHANGED empty, so the classifier fails open and every
-        # lane runs. Post-merge / on-demand validation is never weakened.
-        if [ "$EVENT_NAME" = "pull_request" ]; then
-          # Use the compare endpoint with the pinned base/head SHAs from the
-          # event payload instead of the "current PR files" endpoint. The SHAs
-          # are frozen at trigger time, so the file list is deterministic even
-          # if the PR receives a new push between trigger and detect.
-          #
-          # Retried: a rate-limit blip or eventual-consistency 404 on a
-          # freshly-pushed HEAD would otherwise silently fall open (all lanes
-          # run — safe, but wasteful and it masks the API failure).
-          #
-          # `.files[]?` (null-safe): with --paginate, a PR more than 100
-          # commits ahead of its merge-base paginates the compare, and pages
-          # after the first carry `files: null` — bare `.files[]` makes jq
-          # die with "cannot iterate over: null", which fails every retry
-          # and forces the fail-open path (seen on stacked PRs). The full
-          # file list (up to the API's 300-file cap) is on page one.
-          CHANGED=""
-          for i in 1 2 3; do
-            if CHANGED="$(gh api \
-              --paginate \
-              "repos/${REPO}/compare/${BASE_SHA}...${HEAD_SHA}" \
-              --jq '.files[]?.filename')"; then
-              break
-            fi
-            if [ "$i" = 3 ]; then
-              echo "::warning::compare API failed after 3 attempts — failing open (all lanes run)"
-              CHANGED=""
-              break
-            fi
-            echo "::warning::compare API failed (attempt $i); retrying in 10s"
-            sleep 10
-          done
-        fi
-
-        echo "Changed files:"
-        printf '%s\n' "${CHANGED:-(none)}"
-        printf '%s\n' "${CHANGED:-}" | python3 scripts/ci/classify_changes.py

.github/actions/get-app-token/action.yml

@@ -1,69 +0,0 @@
-name: Get GitHub App Token
-description: >-
-  Mint a short-lived (1-hour) installation access token from the repo's
-  GitHub App, replacing the long-lived AUTOFIX_BOT_PAT. App tokens get
-  5,000 req/hr per installation (vs 1,000 for the default GITHUB_TOKEN)
-  and are scoped to the App's installation permissions, not a user account.
-
-  Callers must source App credentials from a protected, main-only environment.
-  Never pass an App private key to a pull_request job, a local action, or a
-  reusable workflow resolved from an untrusted PR ref. The fallback keeps a
-  trusted caller functional when its protected environment is misconfigured.
-
-  Composite actions cannot access contexts directly, so callers pass the
-  public vars.APP_CLIENT_ID and protected secrets.APP_PRIVATE_KEY as inputs.
-  When the private key is empty, the fallback fires.
-
-inputs:
-  client-id:
-    description: GitHub App Client ID. Pass vars.APP_CLIENT_ID from the calling workflow.
-    required: false
-    default: ''
-  private-key:
-    description: GitHub App private key PEM. Pass secrets.APP_PRIVATE_KEY from the calling workflow.
-    required: false
-    default: ''
-  owner:
-    description: GitHub App installation owner. Empty scopes the token to the current repository.
-    required: false
-    default: ''
-  repositories:
-    description: Comma- or newline-separated repositories to scope within the installation owner.
-    required: false
-    default: ''
-
-outputs:
-  token:
-    description: A GitHub App installation access token (1-hour TTL), or GITHUB_TOKEN on forks.
-    value: ${{ steps.app-token.outputs.token || steps.fallback.outputs.token }}
-
-runs:
-  using: composite
-  steps:
-    - name: Check if App credentials exist
-      id: check
-      shell: bash
-      env:
-        CLIENT_ID: ${{ inputs.client-id }}
-      run: |
-        if [ -n "$CLIENT_ID" ]; then
-          echo "has_app=true" >> "$GITHUB_OUTPUT"
-        else
-          echo "has_app=false" >> "$GITHUB_OUTPUT"
-        fi
-
-    - name: Create GitHub App token
-      id: app-token
-      if: steps.check.outputs.has_app == 'true'
-      uses: actions/create-github-app-token@bcd2ba49218906704ab6c1aa796996da409d3eb1 # v3.2.0
-      with:
-        client-id: ${{ inputs.client-id }}
-        private-key: ${{ inputs.private-key }}
-        owner: ${{ inputs.owner }}
-        repositories: ${{ inputs.repositories }}
-
-    - name: Fall back to GITHUB_TOKEN
-      id: fallback
-      if: steps.check.outputs.has_app != 'true'
-      shell: bash
-      run: echo "token=${{ github.token }}" >> "$GITHUB_OUTPUT"

.github/actions/hermes-smoke-test/action.yml

@@ -0,0 +1,50 @@
+name: Hermes smoke test
+description: >
+  Run the image's built-in entrypoint against `--help` and `dashboard --help`
+  to catch basic runtime regressions before publishing.  Requires the image
+  to already be loaded into the local Docker daemon under `image`.
+
+  Works identically on amd64 and arm64 runners.
+
+inputs:
+  image:
+    description: Fully-qualified image tag (e.g. nousresearch/hermes-agent:test)
+    required: true
+
+runs:
+  using: composite
+  steps:
+    - name: Ensure /tmp/hermes-test is hermes-writable
+      shell: bash
+      run: |
+        # The image runs as the hermes user (UID 10000).  GitHub Actions
+        # creates /tmp/hermes-test root-owned by default, which hermes
+        # can't write to — chown it to match the in-container UID before
+        # bind-mounting.  Real users doing `docker run -v ~/.hermes:...`
+        # with their own UID hit the same issue and have their own
+        # remediations (HERMES_UID env var, or chown locally).
+        mkdir -p /tmp/hermes-test
+        sudo chown -R 10000:10000 /tmp/hermes-test
+
+    - name: hermes --help
+      shell: bash
+      run: |
+        # Use the image's real ENTRYPOINT (/init + main-wrapper.sh) so
+        # this exercises the actual production startup path. PR #30136
+        # review caught that an --entrypoint override here had been
+        # silently neutered by the s6-overlay migration — stage2-hook
+        # ignores its CMD args, so the smoke test was a no-op.
+        docker run --rm \
+          -v /tmp/hermes-test:/opt/data \
+          "${{ inputs.image }}" --help
+
+    - name: hermes dashboard --help
+      shell: bash
+      run: |
+        # Regression guard for #9153: dashboard was present in source but
+        # missing from the published image.  If this fails, something in
+        # the Dockerfile is excluding the dashboard subcommand from the
+        # installed package.
+        docker run --rm \
+          -v /tmp/hermes-test:/opt/data \
+          "${{ inputs.image }}" dashboard --help

.github/actions/retry/action.yml

@@ -1,70 +0,0 @@
-name: Retry a flaky command
-description: >-
-  Run a shell command, retrying on non-zero exit. For dependency installs
-  (npm ci, uv sync) whose only failures are transient network/toolchain
-  flakes — a node-gyp header fetch, a registry blip — so CI self-heals
-  instead of needing a manual re-run. Can also capture stdout as a step
-  output for commands whose result must be consumed by later steps.
-
-inputs:
-  command:
-    description: Shell command to run (and retry).
-    required: true
-  attempts:
-    description: Max attempts before giving up.
-    default: "3"
-  delay:
-    description: Seconds to wait between attempts.
-    default: "10"
-  working-directory:
-    description: Directory to run in.
-    default: "."
-
-outputs:
-  stdout:
-    description: Captured stdout from the successful attempt (empty if not needed).
-    value: ${{ steps.retry.outputs.stdout }}
-
-runs:
-  using: composite
-  steps:
-    - id: retry
-      shell: bash
-      working-directory: ${{ inputs.working-directory }}
-      # command goes through env, never interpolated into the script body, so
-      # a command with quotes/specials can't break or inject into the runner.
-      env:
-        _CMD: ${{ inputs.command }}
-        _ATTEMPTS: ${{ inputs.attempts }}
-        _DELAY: ${{ inputs.delay }}
-      run: |
-        set -uo pipefail
-        _OUTFILE="$(mktemp)"
-        trap 'rm -f "$_OUTFILE"' EXIT
-        n=0
-        while :; do
-          n=$((n + 1))
-          echo "::group::attempt $n/$_ATTEMPTS: $_CMD"
-          # Run the command, capturing stdout to a temp file while still
-          # streaming to the log. We redirect first, then tee the file to
-          # stdout — this avoids pipefail + tee exit-code interactions that
-          # can cause the if-branch to be skipped under set -e.
-          if bash -c "$_CMD" > "$_OUTFILE"; then
-            cat "$_OUTFILE"
-            echo "::endgroup::"
-            # Preserve newlines in the output via heredoc delimiter.
-            {
-              echo 'stdout<<__RETRY_STDOUT_EOF__'
-              cat "$_OUTFILE"
-              echo '__RETRY_STDOUT_EOF__'
-            } >> "$GITHUB_OUTPUT"
-            exit 0
-          fi
-          echo "::endgroup::"
-          if [ "$n" -ge "$_ATTEMPTS" ]; then
-            echo "::error::failed after $n attempts: $_CMD"
-            exit 1
-          fi
-          echo "::warning::attempt $n failed; retrying in ${_DELAY}s: $_CMD"
-          sleep "$_DELAY"
-        done

.github/workflows/build-windows-installer.yml

@@ -0,0 +1,100 @@
+name: Build Windows Installer
+
+on:
+  workflow_dispatch:
+
+permissions:
+  contents: read
+
+jobs:
+  # Gate: workflow_dispatch is already restricted to users with write access,
+  # but we want ADMIN-only. Explicitly check the triggering actor's repo
+  # permission via the API and fail fast for anyone below admin.
+  authorize:
+    name: Authorize (admins only)
+    runs-on: ubuntu-latest
+    timeout-minutes: 5
+    steps:
+      - name: Check actor is a repo admin
+        env:
+          GH_TOKEN: ${{ github.token }}
+          ACTOR: ${{ github.actor }}
+        run: |
+          set -euo pipefail
+          perm=$(gh api \
+            "repos/${{ github.repository }}/collaborators/${ACTOR}/permission" \
+            --jq '.permission')
+          echo "Actor '${ACTOR}' has permission: ${perm}"
+          if [ "${perm}" != "admin" ]; then
+            echo "::error::'${ACTOR}' is not a repo admin (permission=${perm}). Refusing to build/sign."
+            exit 1
+          fi
+          echo "Authorized: '${ACTOR}' is an admin."
+
+  build:
+    name: Hermes-Setup.exe
+    needs: authorize
+    runs-on: windows-latest
+    timeout-minutes: 30
+    permissions:
+      contents: read
+      # Required for OIDC auth to Azure (azure/login federated credentials).
+      id-token: write
+
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd  # v6.0.2
+
+      - name: Setup Node.js
+        uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020  # v4
+        with:
+          node-version: 22
+          cache: npm
+
+      - name: Install npm dependencies
+        run: npm ci
+
+      - name: Setup Rust
+        uses: dtolnay/rust-toolchain@29eef336d9b2848a0b548edc03f92a220660cdb8  # stable
+
+      - name: Cache Rust targets
+        uses: Swatinem/rust-cache@e18b497796c12c097a38f9edb9d0641fb99eee32  # v2
+        with:
+          workspaces: apps/bootstrap-installer/src-tauri
+
+      - name: Build installer
+        run: npm run tauri:build
+        working-directory: apps/bootstrap-installer
+
+      - name: Azure login (OIDC)
+        uses: azure/login@a457da9ea143d694b1b9c7c869ebb04ebe844ef5  # v2
+        with:
+          client-id: ${{ secrets.AZURE_CLIENT_ID }}
+          tenant-id: ${{ secrets.AZURE_TENANT_ID }}
+          subscription-id: ${{ secrets.AZURE_SUBSCRIPTION_ID }}
+
+      - name: Sign Hermes-Setup.exe with Azure Artifact Signing
+        uses: azure/artifact-signing-action@c7ab2a863ab5f9a846ddb8265964877ef296ee82  # v2
+        with:
+          endpoint: ${{ vars.AZURE_SIGNING_ENDPOINT }}
+          signing-account-name: ${{ vars.AZURE_SIGNING_ACCOUNT_NAME }}
+          certificate-profile-name: ${{ vars.AZURE_SIGNING_CERTIFICATE_PROFILE }}
+          # Sign both the raw exe and the bundled NSIS installer.
+          files-folder: ${{ github.workspace }}\apps\bootstrap-installer\src-tauri\target\release
+          files-folder-filter: exe
+          files-folder-recurse: true
+          file-digest: SHA256
+          timestamp-rfc3161: http://timestamp.acs.microsoft.com
+          timestamp-digest: SHA256
+
+      - name: Upload NSIS installer
+        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a  # v7.0.1
+        with:
+          name: Hermes-Setup-installer
+          path: apps/bootstrap-installer/src-tauri/target/release/bundle/nsis/*.exe
+
+      - name: Upload raw exe
+        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a  # v7.0.1
+        with:
+          name: Hermes-Setup-exe
+          path: apps/bootstrap-installer/src-tauri/target/release/Hermes-Setup.exe

.github/workflows/ci.yml

@@ -1,397 +0,0 @@
-name: CI
-
-# Orchestrator workflow. Runs ``detect-changes`` once, then conditionally
-# calls the sub-workflows that a PR can actually affect. A final
-# ``all-checks-pass`` gate job aggregates results so branch protection only
-# needs to require a single check.
-#
-# Sub-workflows are triggered via ``workflow_call`` and keep their own job
-# definitions, matrices, and concurrency settings. They no longer have
-# ``push:`` / ``pull_request:`` triggers of their own — everything flows
-# through this file.
-#
-# SECURITY: this workflow runs PR-controlled actions, workflows, and code.
-# Do not add ``secrets: inherit`` or GitHub App credentials here. Trusted
-# main-only automation uses protected environments in its own workflows.
-
-on:
-  pull_request:
-  push:
-    branches: [main]
-
-permissions:
-  contents: read
-  pull-requests: write # needed by lint (PR comment) + supply-chain review_status
-  actions: read # needed by osv-scanner (SARIF upload)
-  security-events: write # needed by osv-scanner (SARIF upload)
-  packages: write # needed by docker build
-
-concurrency:
-  group: ci-${{ github.ref }}
-  cancel-in-progress: ${{ github.event_name == 'pull_request' }}
-
-jobs:
-  # ─────────────────────────────────────────────────────────────────────
-  # detect: run the classifier once. Every downstream job reads its outputs
-  # to decide whether to run. On push/dispatch the classifier fails open
-  # (all lanes true) so post-merge validation is never weakened.
-  # ─────────────────────────────────────────────────────────────────────
-  detect:
-    name: Detect affected areas
-    runs-on: ubuntu-latest
-    timeout-minutes: 10
-    outputs:
-      python: ${{ steps.classify.outputs.python }}
-      frontend: ${{ steps.classify.outputs.frontend }}
-      site: ${{ steps.classify.outputs.site }}
-      scan: ${{ steps.classify.outputs.scan }}
-      deps: ${{ steps.classify.outputs.deps }}
-      npm_lock: ${{ steps.classify.outputs.npm_lock }}
-      docker_meta: ${{ steps.classify.outputs.docker_meta }}
-      mcp_catalog: ${{ steps.classify.outputs.mcp_catalog }}
-      ci_review: ${{ steps.classify.outputs.ci_review }}
-      ci_review_files: ${{ steps.classify.outputs.ci_review_files }}
-      event_name: ${{ github.event_name }}
-    steps:
-      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
-      - name: Detect affected areas
-        id: classify
-        uses: ./.github/actions/detect-changes
-        with:
-          github-token: ${{ github.token }}
-
-  # ─────────────────────────────────────────────────────────────────────
-  # Lane-gated sub-workflows. Each runs in parallel after detect finishes.
-  # Skipped workflows (if condition is false) don't spin up runners.
-  # ─────────────────────────────────────────────────────────────────────
-  tests:
-    name: Python tests
-    needs: detect
-    if: needs.detect.outputs.python == 'true'
-    uses: ./.github/workflows/tests.yml
-    with:
-      slice_count: 8
-
-  lint:
-    name: Python lints
-    needs: detect
-    if: needs.detect.outputs.python == 'true'
-    uses: ./.github/workflows/lint.yml
-    with:
-      event_name: ${{ needs.detect.outputs.event_name }}
-
-  js-tests:
-    name: JS & TS checks
-    needs: detect
-    if: needs.detect.outputs.frontend == 'true'
-    uses: ./.github/workflows/js-tests.yml
-
-  e2e-desktop:
-    name: Desktop E2E
-    needs: detect
-    if: needs.detect.outputs.python == 'true' || needs.detect.outputs.frontend == 'true'
-    uses: ./.github/workflows/e2e-desktop.yml
-
-  docs-site:
-    name: Docs Site
-    needs: detect
-    if: needs.detect.outputs.site == 'true'
-    uses: ./.github/workflows/docs-site-checks.yml
-
-  history-check:
-    name: Deny unrelated histories
-    needs: detect
-    if: needs.detect.outputs.event_name == 'pull_request'
-    uses: ./.github/workflows/history-check.yml
-
-  contributor-check:
-    name: Check contributors
-    needs: detect
-    if: needs.detect.outputs.python == 'true'
-    uses: ./.github/workflows/contributor-check.yml
-
-  uv-lockfile:
-    name: Check uv.lock
-    needs: detect
-    uses: ./.github/workflows/uv-lockfile-check.yml
-
-  lockfile-diff:
-    name: package-lock.json diff
-    needs: detect
-    if: needs.detect.outputs.event_name == 'pull_request' && needs.detect.outputs.npm_lock == 'true'
-    uses: ./.github/workflows/lockfile-diff.yml
-
-  docker-lint:
-    name: Lint Docker scripts
-    needs: detect
-    if: needs.detect.outputs.docker_meta == 'true'
-    uses: ./.github/workflows/docker-lint.yml
-
-  docker:
-    name: Build&Test Docker image
-    needs: detect
-    # Trusted main pushes run docker.yml directly so its container-publish
-    # environment secrets never cross this reusable-workflow call. PR runs
-    # remain build/test-only and secret-free.
-    if: needs.detect.outputs.event_name == 'pull_request' && (needs.detect.outputs.python == 'true' || needs.detect.outputs.frontend == 'true' || needs.detect.outputs.docker_meta == 'true')
-    uses: ./.github/workflows/docker.yml
-
-  supply-chain:
-    name: Supply-chain scan
-    needs: detect
-    if: needs.detect.outputs.event_name == 'pull_request' && (needs.detect.outputs.scan == 'true' || needs.detect.outputs.deps == 'true')
-    uses: ./.github/workflows/supply-chain-audit.yml
-    with:
-      event_name: ${{ needs.detect.outputs.event_name }}
-      scan: ${{ needs.detect.outputs.scan == 'true' }}
-      deps: ${{ needs.detect.outputs.deps == 'true' }}
-
-  review-labels:
-    name: Review label gate
-    needs: [detect, supply-chain]
-    if: always() && needs.detect.outputs.event_name == 'pull_request' && (needs.detect.outputs.ci_review == 'true' || needs.detect.outputs.mcp_catalog == 'true' || needs.supply-chain.outputs.critical_findings == 'true')
-    uses: ./.github/workflows/review-labels.yml
-    with:
-      ci_review: ${{ needs.detect.outputs.ci_review == 'true' }}
-      ci_review_files: ${{ needs.detect.outputs.ci_review_files }}
-      mcp_catalog: ${{ needs.detect.outputs.mcp_catalog == 'true' }}
-      supply_chain: ${{ needs.supply-chain.outputs.critical_findings == 'true' }}
-
-  osv-scanner:
-    name: OSV scan
-    uses: ./.github/workflows/osv-scanner.yml
-
-  # ─────────────────────────────────────────────────────────────────────
-  # Live-updating PR review comment.
-  #
-  # A single ``comment-live`` job polls the GitHub Actions API every 15s
-  # for job statuses in this run, re-assembles the review comment from
-  # whatever results are available, and upserts it via the
-  # ``<!-- hermes-ci-review-bot -->`` marker.
-  #
-  # When the visible job set goes quiet, the poller waits 10 seconds and polls
-  # once more so downstream jobs created by an aggregate gate get included.
-  # ─────────────────────────────────────────────────────────────────────
-  comment-live:
-    name: CI review comment (live)
-    needs: [detect, review-labels, lockfile-diff, supply-chain, osv-scanner, uv-lockfile, history-check, contributor-check, e2e-desktop]
-    if: always() && github.event_name == 'pull_request' && github.event.pull_request.head.repo.fork != true
-    runs-on: ubuntu-latest
-    timeout-minutes: 40
-    steps:
-      - name: Checkout code
-        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
-        with:
-          ref: ${{ github.event.repository.default_branch }}
-          persist-credentials: false
-
-      - name: Run live comment poller
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-          GITHUB_REPOSITORY: ${{ github.repository }}
-          GITHUB_RUN_ID: ${{ github.run_id }}
-          PR_NUMBER: ${{ github.event.pull_request.number }}
-          RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
-          # Commit info for the review comment header.
-          COMMIT_SHA: ${{ github.event.pull_request.head.sha }}
-          COMMIT_MESSAGE: ${{ github.event.pull_request.head.commit.message }}
-          COMMIT_URL: ${{ github.server_url }}/${{ github.repository }}/pull/${{ github.event.pull_request.number }}/commits/${{ github.event.pull_request.head.sha }}
-          # Structured review statuses from workflow_call jobs.
-          # Each job outputs a JSON array of {source, results: [...]} objects
-          # that the assembler renders directly — no hardcoded job-name
-          # matching. We merge all available outputs into one array.
-          REVIEW_STATUSES: ${{ toJSON(needs.*.outputs.review_status) }}
-        run: |
-          set -uo pipefail
-
-          # REVIEW_STATUSES is a JSON array of strings (some may be empty
-          # when a job was skipped). Parse each string and merge into one
-          # flat array for the assembler.
-          python3 - <<'PYEOF'
-          import json, os, sys
-
-          raw = os.environ.get("REVIEW_STATUSES", "")
-          merged = []
-          if raw:
-              try:
-                  arr = json.loads(raw)
-              except (json.JSONDecodeError, TypeError):
-                  arr = []
-              for item in arr:
-                  if not item:
-                      continue
-                  try:
-                      statuses = json.loads(item)
-                  except (json.JSONDecodeError, TypeError):
-                      continue
-                  if isinstance(statuses, list):
-                      merged.extend(statuses)
-
-          # Write merged array to a temp file the poller reads.
-          with open("/tmp/review_statuses.json", "w") as f:
-              json.dump(merged, f)
-          print(f"Merged {len(merged)} review status entries")
-          PYEOF
-
-          python3 scripts/ci/live_comment.py \
-            --interval 15 \
-            --timeout 2100 \
-            --review-statuses-file /tmp/review_statuses.json
-
-  # ─────────────────────────────────────────────────────────────────────
-  # Gate: runs after everything. ``if: always()`` ensures it reports a
-  # status even when some deps were skipped. Only actual ``failure``
-  # results cause it to fail; ``skipped`` is treated as success.
-  #
-  # Branch protection should require ONLY this check.
-  #
-  # Outputs ``needs-json`` — a compact ``{job_name: result}`` dict — so
-  # the live comment poller can list failed jobs in the PR comment.
-  # ─────────────────────────────────────────────────────────────────────
-  all-checks-pass:
-    name: All required checks pass
-    needs:
-      - detect
-      - tests
-      - lint
-      - js-tests
-      - e2e-desktop
-      - docs-site
-      - history-check
-      - contributor-check
-      - uv-lockfile
-      - lockfile-diff
-      - docker-lint
-      - supply-chain
-      - review-labels
-      - osv-scanner
-      # comment-live is a polling job — it doesn't block the gate.
-      # we don't require docker to pass rn because it's so slow lol
-      # - docker
-    if: always()
-    runs-on: ubuntu-latest
-    timeout-minutes: 10
-    outputs:
-      needs-json: ${{ steps.evaluate.outputs.needs-json }}
-    steps:
-      - name: Evaluate job results
-        id: evaluate
-        env:
-          NEEDS: ${{ toJSON(needs) }}
-        run: |
-          echo "$NEEDS" | python3 -c "
-          import json, sys
-          needs = json.load(sys.stdin)
-          # Emit compact {job_name: result} for the comment assembler.
-          compact = {name: info['result'] for name, info in needs.items()}
-          print(f'needs-json={json.dumps(compact)}')
-          with open('$GITHUB_OUTPUT', 'a') as f:
-              f.write(f'needs-json={json.dumps(compact)}\n')
-          failed = [name for name, info in needs.items() if info['result'] == 'failure']
-          for name, info in sorted(needs.items()):
-              result = info['result']
-              icon = '✅' if result in ('success', 'skipped') else '❌'
-              print(f'{icon} {name}: {result}')
-          if failed:
-              print(f'::error::{len(failed)} job(s) failed: {\", \".join(failed)}')
-              sys.exit(1)
-          print('All checks passed (or were skipped)')
-          "
-
-  # ─────────────────────────────────────────────────────────────────────
-  # CI timing report: collect per-job/step durations from the GitHub API,
-  # cache them on main (as a baseline), and on PRs generate an HTML diff
-  # report with a gantt chart + per-step breakdown. The report is uploaded
-  # as an artifact and a markdown summary is written to $GITHUB_STEP_SUMMARY.
-  #
-  # The live comment poller can read the standalone review-status artifact
-  # after the HTML report is uploaded, so its link points straight at that report.
-  # ─────────────────────────────────────────────────────────────────────
-  ci-timings:
-    name: CI timing report
-    needs: [all-checks-pass, docker]
-    if: always()
-    runs-on: ubuntu-latest
-    timeout-minutes: 10
-    steps:
-      - name: Checkout code
-        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
-
-      - name: Restore baseline cache (PR only)
-        if: github.event_name == 'pull_request'
-        uses: actions/cache/restore@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
-        with:
-          path: ci-timings-baseline.json
-          # Prefix-match: exact key will never hit (run_id differs), so
-          # restore-keys finds the most recent baseline from main.
-          key: ci-timings-baseline-never-exact
-          restore-keys: |
-            ci-timings-baseline-
-
-      - name: Collect timings and generate report
-        env:
-          GITHUB_TOKEN: ${{ github.token }}
-        run: |
-          python3 scripts/ci/timings_report.py \
-            --baseline ci-timings-baseline.json \
-            --output ci-timings-report.html \
-            --json-out ci-timings.json \
-            --summary-out ci-timings-summary.md
-
-      - name: Upload HTML report
-        # Advisory report — artifact-service blips must not fail the job.
-        continue-on-error: true
-        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
-        id: ci-timings-html
-        with:
-          name: ci-timings-report
-          path: ci-timings-report.html
-          retention-days: 14
-
-      - name: Build linked review status
-        if: hashFiles('ci-timings.json') != ''
-        env:
-          CI_TIMINGS_REPORT_URL: ${{ steps.ci-timings-html.outputs.artifact-url }}
-        run: |
-          python3 scripts/ci/timings_report.py \
-            --from-json ci-timings.json \
-            --baseline ci-timings-baseline.json \
-            --review-status-out review-status.json \
-            --review-status-only
-
-      - name: Upload review status
-        if: hashFiles('review-status.json') != ''
-        continue-on-error: true
-        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
-        with:
-          name: ci-timings-review-status
-          path: review-status.json
-          retention-days: 14
-
-      - name: Output summary
-        env:
-          REPORT_URL: ${{ steps.ci-timings-html.outputs.artifact-url}}
-        run: |
-          {
-            echo "# CI Timing report"
-            echo "[View the full interactive report]($REPORT_URL)"
-          } >> "$GITHUB_STEP_SUMMARY"
-          cat ci-timings-summary.md >> "$GITHUB_STEP_SUMMARY"
-
-      - name: Save baseline cache (main only)
-        if: github.event_name == 'push' && github.ref == 'refs/heads/main'
-        run: |
-          # Degraded runs (API rate-limited) produce no ci-timings.json —
-          # skip rather than fail, and never cache an empty baseline.
-          if [ -f ci-timings.json ]; then
-            cp ci-timings.json ci-timings-baseline.json
-          else
-            echo "No timings JSON this run — skipping baseline update"
-          fi
-
-      - name: Upload baseline to cache (main only)
-        if: github.event_name == 'push' && github.ref == 'refs/heads/main' && hashFiles('ci-timings-baseline.json') != ''
-        uses: actions/cache/save@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
-        with:
-          path: ci-timings-baseline.json
-          key: ci-timings-baseline-${{ github.run_id }}

.github/workflows/contributor-check.yml

@@ -1,28 +1,37 @@
 name: Contributor Attribution Check
 
 on:
-  workflow_call:
-    outputs:
-      review_status:
-        description: "JSON array of review status objects"
-        value: ${{ jobs.check-attribution.outputs.review_status }}
-
+  # No paths filter — the job must always run so the required check
+  # reports a status (path-gated workflows leave checks "pending" forever
+  # when no matching files change, which blocks merge).
+  pull_request:
+    branches: [main]
 permissions:
   contents: read
 
 jobs:
   check-attribution:
     runs-on: ubuntu-latest
-    timeout-minutes: 10
-    outputs:
-      review_status: ${{ steps.check-emails.outputs.review_status }}
     steps:
       - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd  # v6.0.2
         with:
           fetch-depth: 0  # Full history needed for git log
 
+      - name: Check if relevant files changed
+        id: filter
+        run: |
+          BASE="${{ github.event.pull_request.base.sha }}"
+          HEAD="${{ github.event.pull_request.head.sha }}"
+          CHANGED=$(git diff --name-only "$BASE"..."$HEAD" -- '*.py' '**/*.py' '.github/workflows/contributor-check.yml' || true)
+          if [ -n "$CHANGED" ]; then
+            echo "run=true" >> "$GITHUB_OUTPUT"
+          else
+            echo "run=false" >> "$GITHUB_OUTPUT"
+            echo "No Python files changed, skipping attribution check."
+          fi
+
       - name: Check for unmapped contributor emails
-        id: check-emails
+        if: steps.filter.outputs.run == 'true'
         run: |
           # Get the merge base between this PR and main
           MERGE_BASE=$(git merge-base origin/main HEAD)
@@ -32,13 +41,10 @@ jobs:
 
           if [ -z "$NEW_EMAILS" ]; then
             echo "No new commits to check."
-            echo "review_status=[]" >> "$GITHUB_OUTPUT"
             exit 0
           fi
 
-          # An email is mapped if it has a file in contributors/emails/
-          # (one file per email — conflict-free) or an entry in the frozen
-          # legacy AUTHOR_MAP in scripts/release.py.
+          # Check each email against AUTHOR_MAP in release.py
           MISSING=""
           while IFS= read -r email; do
             # Skip teknium and bot emails
@@ -47,12 +53,9 @@ jobs:
                 continue ;;
             esac
 
+            # Check if email is in AUTHOR_MAP (either as a key or matches noreply pattern)
             if echo "$email" | grep -qP '\+.*@users\.noreply\.github\.com'; then
-              continue  # GitHub id+login noreply emails auto-resolve
-            fi
-
-            if [ -f "contributors/emails/${email}" ]; then
-              continue  # mapped via the contributors directory
+              continue  # GitHub noreply emails auto-resolve
             fi
 
             if ! grep -qF "\"${email}\"" scripts/release.py 2>/dev/null; then
@@ -63,29 +66,19 @@ jobs:
 
           if [ -n "$MISSING" ]; then
             echo ""
-            echo "⚠️  New contributor email(s) without a mapping:"
+            echo "⚠️  New contributor email(s) not in AUTHOR_MAP:"
             echo -e "$MISSING"
             echo ""
-            echo "Add a mapping file (do NOT edit AUTHOR_MAP in release.py):"
+            echo "Please add mappings to scripts/release.py AUTHOR_MAP:"
             echo -e "$MISSING" | while read -r line; do
               email=$(echo "$line" | sed 's/^ *//' | cut -d' ' -f1)
               [ -z "$email" ] && continue
-              echo "    python3 scripts/add_contributor.py ${email} <github-username>"
+              echo "    \"${email}\": \"<github-username>\","
             done
             echo ""
             echo "To find the GitHub username for an email:"
             echo "  gh api 'search/users?q=EMAIL+in:email' --jq '.items[0].login'"
-
-            # Emit review_status for unmapped emails
-            DETAIL=$(echo -e "$MISSING" | sed '/^$/d; s/^  //')
-            HOW_TO_FIX=$'Add mappings to scripts/release.py AUTHOR_MAP:\n```\n"<email>": "<github-username>",\n```\nTo find the GitHub username for an email:\n```\ngh api \'search/users?q=EMAIL+in:email\' --jq \'.items[0].login\'\n```\n'
-            REVIEW_STATUS=$(jq -nc \
-              --arg detail "$DETAIL" \
-              --arg how_to_fix "$HOW_TO_FIX" \
-              '[{"source":"contributor attribution","results":[{"kind":"action_required","title":"Unmapped contributor email(s)","summary":"New contributor email(s) are not in AUTHOR_MAP.","detail":$detail,"how_to_fix":$how_to_fix}]}]')
-            echo "review_status=$REVIEW_STATUS" >> "$GITHUB_OUTPUT"
-
             exit 1
           else
-            echo "✅ All contributor emails are mapped."
+            echo "✅ All contributor emails are mapped in AUTHOR_MAP."
           fi

.github/workflows/deploy-site.yml

@@ -41,28 +41,19 @@ jobs:
     # doesn't auto-deploy via the deploy-docs path.
     if: github.event_name == 'release' || github.event_name == 'workflow_dispatch'
     runs-on: ubuntu-latest
-    timeout-minutes: 30
     steps:
       - name: Trigger Vercel Deploy
-        run: curl -fsS --retry 3 --retry-delay 10 -X POST "${{ secrets.VERCEL_DEPLOY_HOOK }}"
+        run: curl -X POST "${{ secrets.VERCEL_DEPLOY_HOOK }}"
 
   deploy-docs:
     if: github.repository == 'NousResearch/hermes-agent'
     runs-on: ubuntu-latest
-    timeout-minutes: 30
     environment:
       name: github-pages
       url: ${{ steps.deploy.outputs.page_url }}
     steps:
       - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd  # v6.0.2
 
-      - name: Get GitHub App token
-        id: app-token
-        uses: ./.github/actions/get-app-token
-        with:
-          client-id: ${{ vars.APP_CLIENT_ID }}
-          private-key: ${{ secrets.APP_PRIVATE_KEY }}
-
       - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020  # v4
         with:
           node-version: 22
@@ -74,14 +65,12 @@ jobs:
           python-version: '3.11'
 
       - name: Install PyYAML for skill extraction
-        uses: ./.github/actions/retry
-        with:
-          command: pip install pyyaml==6.0.2 httpx==0.28.1
+        run: pip install pyyaml==6.0.2 httpx==0.28.1
 
       - name: Prepare skills index (unified multi-source catalog)
         env:
-          GH_TOKEN: ${{ steps.app-token.outputs.token }}
-          GITHUB_TOKEN: ${{ steps.app-token.outputs.token }}
+          GH_TOKEN: ${{ github.token }}
+          GITHUB_TOKEN: ${{ github.token }}
           SKILLS_INDEX_RUN_ID: ${{ github.event.inputs.skills_index_run_id || '' }}
           REBUILD_SKILLS_INDEX: ${{ github.event.inputs.rebuild_skills_index || 'false' }}
         run: |
@@ -161,10 +150,8 @@ jobs:
         run: python3 website/scripts/generate-skill-docs.py
 
       - name: Install dependencies
-        uses: ./.github/actions/retry
-        with:
-          command: npm ci
-          working-directory: website
+        run: npm ci
+        working-directory: website
 
       - name: Build Docusaurus
         run: npm run build

.github/workflows/docker-lint.yml

@@ -2,7 +2,7 @@ name: Docker / shell lint
 
 # Lints the container build inputs: Dockerfile (via hadolint) and any shell
 # scripts under docker/ (via shellcheck). These catch the class of regression
-# the behavioral docker smoke test can't — unquoted variable
+# the behavioral docker-publish smoke test can't — unquoted variable
 # expansions, silently-failing RUN commands, etc.
 #
 # Rules and ignores are documented in .hadolint.yaml at the repo root.
@@ -11,7 +11,19 @@ name: Docker / shell lint
 # activate script doesn't exist at lint time.
 
 on:
-  workflow_call:
+  push:
+    branches: [main]
+    paths:
+      - Dockerfile
+      - docker/**
+      - .hadolint.yaml
+      - .github/workflows/docker-lint.yml
+
+  # No paths filter — the job must always run so the required check
+  # reports a status (path-gated workflows leave checks "pending" forever
+  # when no matching files change, which blocks merge).
+  pull_request:
+    branches: [main]
 
 permissions:
   contents: read

.github/workflows/docker-publish.yml

@@ -0,0 +1,357 @@
+name: Docker Build and Publish
+
+on:
+  push:
+    branches: [main]
+    paths:
+      - '**/*.py'
+      - 'pyproject.toml'
+      - 'uv.lock'
+      - 'Dockerfile'
+      - 'docker/**'
+      - '.github/workflows/docker-publish.yml'
+      - '.github/actions/hermes-smoke-test/**'
+
+  # No paths filter — the job must always run so the required check
+  # reports a status (path-gated workflows leave checks "pending" forever
+  # when no matching files change, which blocks merge).
+  pull_request:
+    branches: [main]
+
+  release:
+    types: [published]
+
+permissions:
+  contents: read
+  # Needed so the arm64 job can push/pull its registry-backed build cache
+  # to ghcr.io (cache-to/cache-from type=registry).  See the build-arm64
+  # job for why registry cache replaced the gha cache on that arch.
+  packages: write
+
+# Concurrency: push/release runs are NEVER cancelled so every merge gets
+# its own image.  PR runs reuse a PR-scoped group with
+# cancel-in-progress: true so rapid pushes to the same PR collapse to the
+# latest commit.
+concurrency:
+  group: docker-${{ github.event.pull_request.number || github.ref }}
+  cancel-in-progress: ${{ github.event_name == 'pull_request' }}
+
+env:
+  IMAGE_NAME: nousresearch/hermes-agent
+
+jobs:
+  # ---------------------------------------------------------------------------
+  # Build amd64 natively.  This job also runs the smoke tests (basic --help
+  # and the dashboard subcommand regression guard from #9153), because amd64
+  # is the only arch we can `load` into the local daemon on an amd64 runner.
+  # ---------------------------------------------------------------------------
+  build-amd64:
+    # Only run on the upstream repository, not on forks
+    if: github.repository == 'NousResearch/hermes-agent'
+    runs-on: ubuntu-latest
+    timeout-minutes: 45
+    outputs:
+      digest: ${{ steps.push.outputs.digest }}
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd  # v6.0.2
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@8d2750c68a42422c14e847fe6c8ac0403b4cbd6f  # v3
+
+      # Build once, load into the local daemon for smoke testing.  Cached
+      # to gha with a per-arch scope; the push step below reuses every
+      # layer from this build.
+      - name: Build image (amd64, smoke test)
+        uses: docker/build-push-action@bcafcacb16a39f128d818304e6c9c0c18556b85f  # v7.1.0
+        with:
+          context: .
+          file: Dockerfile
+          load: true
+          platforms: linux/amd64
+          tags: ${{ env.IMAGE_NAME }}:test
+          build-args: |
+            HERMES_GIT_SHA=${{ github.sha }}
+          cache-from: type=gha,scope=docker-amd64
+          cache-to: type=gha,mode=max,scope=docker-amd64
+
+      - name: Smoke test image
+        uses: ./.github/actions/hermes-smoke-test
+        with:
+          image: ${{ env.IMAGE_NAME }}:test
+
+      # ---------------------------------------------------------------------
+      # Run the docker-integration test suite against the freshly-built
+      # image already loaded into the local daemon (`:test`).  These tests
+      # are excluded from the sharded `tests.yml :: test` matrix on purpose
+      # (see `_SKIP_PARTS` in scripts/run_tests_parallel.py) because each
+      # shard would otherwise reach the session-scoped ``built_image``
+      # fixture in ``tests/docker/conftest.py`` and start a 3-7min
+      # ``docker build`` — guaranteed to
+      # die in fixture setup.
+      #
+      # Piggybacking here avoids a second image build: the smoke test
+      # already proved the image loads + runs, so the daemon has it under
+      # `${IMAGE_NAME}:test` and we just point ``HERMES_TEST_IMAGE`` at
+      # that.  The fixture's ``HERMES_TEST_IMAGE`` branch (see
+      # tests/docker/conftest.py:62-63) short-circuits the rebuild.
+      #
+      # Why this job and not a standalone one: the image is 5GB+; passing
+      # it between jobs via ``docker save``/``upload-artifact`` is slower
+      # than the build itself.  Reusing the existing daemon state is the
+      # cheapest path to coverage on every PR that touches docker code.
+      # ---------------------------------------------------------------------
+      - name: Install uv (for docker tests)
+        uses: astral-sh/setup-uv@d4b2f3b6ecc6e67c4457f6d3e41ec42d3d0fcb86  # v5
+
+      - name: Set up Python 3.11 (for docker tests)
+        run: uv python install 3.11
+
+      - name: Install Python dependencies (for docker tests)
+        run: |
+          uv venv .venv --python 3.11
+          source .venv/bin/activate
+          # ``dev`` extra pulls in pytest, pytest-asyncio —
+          # everything tests/docker/ needs.  We deliberately avoid ``all``
+          # here because the docker tests only drive the container via
+          # subprocess and don't import hermes_agent's optional deps.
+          uv pip install -e ".[dev]"
+
+      - name: Run docker integration tests
+        env:
+          # Skip rebuild; use the image already loaded by the build step.
+          HERMES_TEST_IMAGE: ${{ env.IMAGE_NAME }}:test
+          # Match the policy in tests.yml :: test job — no accidental
+          # real-API calls from inside the harness.
+          OPENROUTER_API_KEY: ""
+          OPENAI_API_KEY: ""
+          NOUS_API_KEY: ""
+        run: |
+          source .venv/bin/activate
+          python -m pytest tests/docker/ -v --tb=short
+
+      - name: Log in to Docker Hub
+        if: github.event_name == 'push' && github.ref == 'refs/heads/main' || github.event_name == 'release'
+        uses: docker/login-action@4907a6ddec9925e35a0a9e82d7399ccc52663121  # v4.1.0
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_TOKEN }}
+
+      # Push amd64 by digest only (no tag).  The merge job assembles the
+      # tagged manifest list.  `push-by-digest=true` is docker's recommended
+      # pattern for multi-runner multi-platform builds.
+      - name: Push amd64 by digest
+        id: push
+        if: github.event_name == 'push' && github.ref == 'refs/heads/main' || github.event_name == 'release'
+        uses: docker/build-push-action@bcafcacb16a39f128d818304e6c9c0c18556b85f  # v7.1.0
+        with:
+          context: .
+          file: Dockerfile
+          platforms: linux/amd64
+          labels: |
+            org.opencontainers.image.revision=${{ github.sha }}
+          build-args: |
+            HERMES_GIT_SHA=${{ github.sha }}
+          outputs: type=image,name=${{ env.IMAGE_NAME }},push-by-digest=true,name-canonical=true,push=true
+          cache-from: type=gha,scope=docker-amd64
+          cache-to: type=gha,mode=max,scope=docker-amd64
+
+      # Write the digest to a file and upload it as an artifact so the
+      # merge job can stitch both per-arch digests into a manifest list.
+      - name: Export digest
+        if: github.event_name == 'push' && github.ref == 'refs/heads/main' || github.event_name == 'release'
+        run: |
+          mkdir -p /tmp/digests
+          digest="${{ steps.push.outputs.digest }}"
+          touch "/tmp/digests/${digest#sha256:}"
+
+      - name: Upload digest artifact
+        if: github.event_name == 'push' && github.ref == 'refs/heads/main' || github.event_name == 'release'
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02  # v4
+        with:
+          name: digest-amd64
+          path: /tmp/digests/*
+          if-no-files-found: error
+          retention-days: 1
+
+  # ---------------------------------------------------------------------------
+  # Build arm64 natively on GitHub's free arm64 runner.  This replaces the
+  # previous QEMU-emulated arm64 build, which was ~5-10x slower and shared
+  # a cache scope with amd64.  Matches the amd64 job's shape: build+load,
+  # smoke test, then on push/release push by digest.
+  # ---------------------------------------------------------------------------
+  build-arm64:
+    if: github.repository == 'NousResearch/hermes-agent'
+    runs-on: ubuntu-24.04-arm
+    timeout-minutes: 45
+    outputs:
+      digest: ${{ steps.push.outputs.digest }}
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd  # v6.0.2
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@8d2750c68a42422c14e847fe6c8ac0403b4cbd6f  # v3
+
+      # Log in to ghcr.io so the registry-backed build cache below can be
+      # read (cache-from) on every event and written (cache-to) on
+      # push/release.  Uses the workflow's GITHUB_TOKEN, which is valid for
+      # the whole job — unlike the gha cache backend's short-lived Azure SAS
+      # token, which expired mid-build on slow cold-cache arm64 runs and
+      # crashed the build before the smoke test (the reason the gha cache
+      # was removed from arm64 PRs in the first place).
+      - name: Log in to ghcr.io (build cache)
+        uses: docker/login-action@4907a6ddec9925e35a0a9e82d7399ccc52663121  # v4.1.0
+        with:
+          registry: ghcr.io
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      # Build once, load into the local daemon for smoke testing.
+      #
+      # PR builds use the registry-backed cache READ-ONLY (cache-from only):
+      # they pull warm layers pushed by the most recent main build but never
+      # write, so rapid PR pushes don't race on cache writes or pollute the
+      # cache ref.  This restores warm-cache speed to arm64 PR builds (which
+      # were running fully uncached and were ~45% slower than amd64, making
+      # them the job most often cancelled on supersede).
+      #
+      # Registry cache (type=registry on ghcr.io) is used instead of the gha
+      # cache that previously broke here: its credential is the job-lifetime
+      # GITHUB_TOKEN, not a short-lived SAS token, so the cold-build-outlives-
+      # token failure mode cannot recur.
+      - name: Build image (arm64, smoke test, cache read-only PR)
+        if: github.event_name == 'pull_request'
+        uses: docker/build-push-action@bcafcacb16a39f128d818304e6c9c0c18556b85f  # v7.1.0
+        with:
+          context: .
+          file: Dockerfile
+          load: true
+          platforms: linux/arm64
+          tags: ${{ env.IMAGE_NAME }}:test
+          build-args: |
+            HERMES_GIT_SHA=${{ github.sha }}
+          cache-from: type=registry,ref=ghcr.io/nousresearch/hermes-agent:buildcache-arm64
+
+      # Main/release builds read AND write the registry cache so the digest
+      # push below reuses layers from this smoke-test build, and so the next
+      # PR/main build starts warm.
+      - name: Build image (arm64, smoke test, cached publish)
+        if: github.event_name != 'pull_request'
+        uses: docker/build-push-action@bcafcacb16a39f128d818304e6c9c0c18556b85f  # v7.1.0
+        with:
+          context: .
+          file: Dockerfile
+          load: true
+          platforms: linux/arm64
+          tags: ${{ env.IMAGE_NAME }}:test
+          build-args: |
+            HERMES_GIT_SHA=${{ github.sha }}
+          cache-from: type=registry,ref=ghcr.io/nousresearch/hermes-agent:buildcache-arm64
+          cache-to: type=registry,ref=ghcr.io/nousresearch/hermes-agent:buildcache-arm64,mode=max
+
+      - name: Smoke test image
+        uses: ./.github/actions/hermes-smoke-test
+        with:
+          image: ${{ env.IMAGE_NAME }}:test
+
+      - name: Log in to Docker Hub
+        if: github.event_name == 'push' && github.ref == 'refs/heads/main' || github.event_name == 'release'
+        uses: docker/login-action@4907a6ddec9925e35a0a9e82d7399ccc52663121  # v4.1.0
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_TOKEN }}
+
+      - name: Push arm64 by digest
+        id: push
+        if: github.event_name == 'push' && github.ref == 'refs/heads/main' || github.event_name == 'release'
+        uses: docker/build-push-action@bcafcacb16a39f128d818304e6c9c0c18556b85f  # v7.1.0
+        with:
+          context: .
+          file: Dockerfile
+          platforms: linux/arm64
+          labels: |
+            org.opencontainers.image.revision=${{ github.sha }}
+          build-args: |
+            HERMES_GIT_SHA=${{ github.sha }}
+          outputs: type=image,name=${{ env.IMAGE_NAME }},push-by-digest=true,name-canonical=true,push=true
+          cache-from: type=registry,ref=ghcr.io/nousresearch/hermes-agent:buildcache-arm64
+          cache-to: type=registry,ref=ghcr.io/nousresearch/hermes-agent:buildcache-arm64,mode=max
+
+      - name: Export digest
+        if: github.event_name == 'push' && github.ref == 'refs/heads/main' || github.event_name == 'release'
+        run: |
+          mkdir -p /tmp/digests
+          digest="${{ steps.push.outputs.digest }}"
+          touch "/tmp/digests/${digest#sha256:}"
+
+      - name: Upload digest artifact
+        if: github.event_name == 'push' && github.ref == 'refs/heads/main' || github.event_name == 'release'
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02  # v4
+        with:
+          name: digest-arm64
+          path: /tmp/digests/*
+          if-no-files-found: error
+          retention-days: 1
+
+  # ---------------------------------------------------------------------------
+  # Stitch both per-arch digests into a single tagged multi-arch manifest.
+  # This is a registry-side operation — no building, no layer re-push —
+  # so it runs in ~30 seconds.
+  #
+  # On main pushes: tags both :main and :latest.
+  # On releases: tags :<release_tag_name>.
+  # ---------------------------------------------------------------------------
+  merge:
+    if: github.repository == 'NousResearch/hermes-agent' && (github.event_name == 'push' && github.ref == 'refs/heads/main' || github.event_name == 'release')
+    runs-on: ubuntu-latest
+    needs: [build-amd64, build-arm64]
+    timeout-minutes: 10
+    steps:
+      - name: Download digests
+        uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093  # v4
+        with:
+          path: /tmp/digests
+          pattern: digest-*
+          merge-multiple: true
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@8d2750c68a42422c14e847fe6c8ac0403b4cbd6f  # v3
+
+      - name: Log in to Docker Hub
+        uses: docker/login-action@4907a6ddec9925e35a0a9e82d7399ccc52663121  # v4.1.0
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_TOKEN }}
+
+      - name: Create manifest list and push
+        working-directory: /tmp/digests
+        run: |
+          set -euo pipefail
+          args=()
+          for digest_file in *; do
+            args+=("${IMAGE_NAME}@sha256:${digest_file}")
+          done
+          if [ "${{ github.event_name }}" = "release" ]; then
+            TAG="${{ github.event.release.tag_name }}"
+            docker buildx imagetools create \
+              -t "${IMAGE_NAME}:${TAG}" \
+              "${args[@]}"
+          else
+            docker buildx imagetools create \
+              -t "${IMAGE_NAME}:main" \
+              -t "${IMAGE_NAME}:latest" \
+              "${args[@]}"
+          fi
+        env:
+          IMAGE_NAME: ${{ env.IMAGE_NAME }}
+
+      - name: Inspect image
+        run: |
+          if [ "${{ github.event_name }}" = "release" ]; then
+            docker buildx imagetools inspect "${IMAGE_NAME}:${{ github.event.release.tag_name }}"
+          else
+            docker buildx imagetools inspect "${IMAGE_NAME}:main"
+          fi
+        env:
+          IMAGE_NAME: ${{ env.IMAGE_NAME }}

.github/workflows/docker.yml

@@ -1,254 +0,0 @@
-name: Docker Build, Test, and Publish
-
-on:
-  # Trusted main pushes run this workflow directly so environment-scoped
-  # Docker Hub secrets are resolved by the top-level workflow, never across
-  # a reusable-workflow boundary.
-  push:
-    branches: [main]
-  release:
-    types: [published]
-  # CI calls this only for untrusted PR build/test coverage. Those runs never
-  # reach the protected publish or merge jobs below.
-  workflow_call:
-
-permissions:
-  contents: read
-
-# Concurrency: push/release runs are NEVER cancelled so every merge gets
-# its own image.  PR runs reuse a PR-scoped group with
-# cancel-in-progress: true so rapid pushes to the same PR collapse to
-# the latest commit.
-concurrency:
-  group: docker-${{ github.event.pull_request.number || github.ref }}
-  cancel-in-progress: ${{ github.event_name == 'pull_request' }}
-
-env:
-  IMAGE_NAME: nousresearch/hermes-agent
-
-jobs:
-  # Build and test the image for each architecture. This job runs PR code,
-  # so it must remain secret-free. Publishing happens in the separate,
-  # protected publish job after these tests pass.
-  build:
-    if: github.repository == 'NousResearch/hermes-agent'
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - arch: amd64
-            runner: ubuntu-latest
-            platform: linux/amd64
-            cache-from: type=gha,scope=docker-amd64
-            cache-to: type=gha,mode=max,scope=docker-amd64
-          - arch: arm64
-            runner: ubuntu-24.04-arm
-            platform: linux/arm64
-            cache-from: type=gha,scope=docker-arm64
-            cache-to: type=gha,mode=max,scope=docker-arm64
-
-    runs-on: ${{ matrix.runner }}
-    timeout-minutes: 45
-    steps:
-      - name: Checkout code
-        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
-
-      - name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@8d2750c68a42422c14e847fe6c8ac0403b4cbd6f # v3
-
-      # Build once, load into the local daemon for testing.  Cached
-      # per-arch; the push step below reuses every layer from this build.
-      - name: Build image (${{ matrix.arch }})
-        uses: docker/build-push-action@bcafcacb16a39f128d818304e6c9c0c18556b85f # v7.1.0
-        with:
-          context: .
-          file: Dockerfile
-          load: true
-          platforms: ${{ matrix.platform }}
-          tags: ${{ env.IMAGE_NAME }}:test
-          build-args: |
-            HERMES_GIT_SHA=${{ github.sha }}
-          cache-from: ${{ matrix.cache-from }}
-          cache-to: ${{ (github.event_name != 'pull_request') && matrix.cache-to || '' }}
-
-
-      # Run the docker-integration test suite against the freshly-built
-      # image already loaded into the local daemon (`:test`).
-      #
-      # Piggybacking here avoids a second image build: the build step
-      # already loaded the image into the daemon under
-      # `${IMAGE_NAME}:test`, so we just point ``HERMES_TEST_IMAGE`` at
-      # that.  The fixture's ``HERMES_TEST_IMAGE`` branch (see
-      # tests/docker/conftest.py:62-63) short-circuits the rebuild.
-      #
-      # Why this job and not a standalone one: the image is 5GB+; passing
-      # it between jobs via ``docker save``/``upload-artifact`` is slower
-      # than the build itself.  Reusing the existing daemon state is the
-      # cheapest path to coverage on every PR that touches docker code.
-      # ---------------------------------------------------------------------
-      - name: Install uv (for docker tests)
-        uses: astral-sh/setup-uv@fac544c07dec837d0ccb6301d7b5580bf5edae39 # 8.2.0
-
-      - name: Set up Python 3.11 (for docker tests)
-        run: uv python install 3.11
-
-      - name: Install Python dependencies (for docker tests)
-        # ``dev`` extra pulls in pytest, pytest-asyncio —
-        # everything tests/docker/ needs.  We deliberately avoid ``all``
-        # here because the docker tests only drive the container via
-        # subprocess and don't import hermes_agent's optional deps.
-        uses: ./.github/actions/retry
-        with:
-          command: uv sync --locked --python 3.11 --extra dev
-
-      - name: Run docker integration tests
-        env:
-          # Skip rebuild; use the image already loaded by the build step.
-          HERMES_TEST_IMAGE: ${{ env.IMAGE_NAME }}:test
-          # Match the policy in tests.yml :: test job — no accidental
-          # real-API calls from inside the harness.
-          OPENROUTER_API_KEY: ""
-          OPENAI_API_KEY: ""
-          NOUS_API_KEY: ""
-        run: |
-          scripts/run_tests.sh tests/docker/ --file-timeout 600
-
-  # ---------------------------------------------------------------------------
-  # Rebuild and push each architecture only after the unprivileged build/test
-  # matrix passes. This job is the sole Docker Hub credential boundary.
-  # ---------------------------------------------------------------------------
-  publish:
-    if: github.repository == 'NousResearch/hermes-agent' && (github.event_name == 'push' && github.ref == 'refs/heads/main' || github.event_name == 'release')
-    needs: [build]
-    environment: container-publish
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - arch: amd64
-            runner: ubuntu-latest
-            platform: linux/amd64
-            cache-from: type=gha,scope=docker-amd64
-            cache-to: type=gha,mode=max,scope=docker-amd64
-          - arch: arm64
-            runner: ubuntu-24.04-arm
-            platform: linux/arm64
-            cache-from: type=gha,scope=docker-arm64
-            cache-to: type=gha,mode=max,scope=docker-arm64
-    runs-on: ${{ matrix.runner }}
-    timeout-minutes: 30
-    steps:
-      - name: Checkout trusted source
-        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
-
-      - name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@8d2750c68a42422c14e847fe6c8ac0403b4cbd6f # v3
-
-      - name: Log in to Docker Hub
-        uses: docker/login-action@4907a6ddec9925e35a0a9e82d7399ccc52663121 # v4.1.0
-        with:
-          username: ${{ secrets.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_TOKEN }}
-
-      # Push by digest only (no tag). The merge job assembles the tagged
-      # manifest list after both architecture publishers complete.
-      - name: Push ${{ matrix.arch }} by digest
-        id: push
-        uses: docker/build-push-action@bcafcacb16a39f128d818304e6c9c0c18556b85f # v7.1.0
-        with:
-          context: .
-          file: Dockerfile
-          platforms: ${{ matrix.platform }}
-          labels: |
-            org.opencontainers.image.revision=${{ github.sha }}
-          build-args: |
-            HERMES_GIT_SHA=${{ github.sha }}
-          outputs: type=image,name=${{ env.IMAGE_NAME }},push-by-digest=true,name-canonical=true,push=true
-          cache-from: ${{ matrix.cache-from }}
-          cache-to: ${{ matrix.cache-to }}
-
-      - name: Export digest
-        run: |
-          mkdir -p /tmp/digests
-          digest="${{ steps.push.outputs.digest }}"
-          touch "/tmp/digests/${digest#sha256:}"
-
-      - name: Upload digest artifact
-        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
-        with:
-          name: digest-${{ matrix.arch }}
-          path: /tmp/digests/*
-          if-no-files-found: error
-          retention-days: 1
-
-  # ---------------------------------------------------------------------------
-  # Stitch both per-arch digests into a single tagged multi-arch manifest.
-  # This is a registry-side operation — no building, no layer re-push —
-  # so it runs in ~30 seconds.
-  #
-  # On main pushes: tags both :main and :latest.
-  # On releases: tags :<release_tag_name>.
-  # ---------------------------------------------------------------------------
-  merge:
-    if: github.repository == 'NousResearch/hermes-agent' && (github.event_name == 'push' && github.ref == 'refs/heads/main' || github.event_name == 'release')
-    runs-on: ubuntu-latest
-    needs: [publish]
-    timeout-minutes: 10
-    environment: container-publish
-    steps:
-      - name: Download digests
-        uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4
-        with:
-          path: /tmp/digests
-          pattern: digest-*
-          merge-multiple: true
-
-      - name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@8d2750c68a42422c14e847fe6c8ac0403b4cbd6f # v3
-
-      - name: Log in to Docker Hub
-        uses: docker/login-action@4907a6ddec9925e35a0a9e82d7399ccc52663121 # v4.1.0
-        with:
-          username: ${{ secrets.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_TOKEN }}
-
-      - name: Create manifest list and push
-        working-directory: /tmp/digests
-        env:
-          IMAGE_NAME: ${{ env.IMAGE_NAME }}
-          RELEASE_TAG: ${{ github.event.release.tag_name }}
-        run: |
-          set -euo pipefail
-          args=()
-          for digest_file in *; do
-            args+=("${IMAGE_NAME}@sha256:${digest_file}")
-          done
-          if [ "${{ github.event_name }}" = "release" ]; then
-            tags=(-t "${IMAGE_NAME}:${RELEASE_TAG}")
-          else
-            tags=(-t "${IMAGE_NAME}:main" -t "${IMAGE_NAME}:latest")
-          fi
-          # Retry: Docker Hub API + just-pushed digest eventual consistency
-          # can transiently fail the create; the operation is idempotent.
-          for i in 1 2 3; do
-            if docker buildx imagetools create "${tags[@]}" "${args[@]}"; then
-              break
-            fi
-            if [ "$i" = 3 ]; then
-              echo "::error::imagetools create failed after 3 attempts"
-              exit 1
-            fi
-            echo "::warning::imagetools create failed (attempt $i); retrying in 20s"
-            sleep 20
-          done
-
-      - name: Inspect image
-        env:
-          IMAGE_NAME: ${{ env.IMAGE_NAME }}
-          RELEASE_TAG: ${{ github.event.release.tag_name }}
-        run: |
-          if [ "${{ github.event_name }}" = "release" ]; then
-            docker buildx imagetools inspect "${IMAGE_NAME}:${RELEASE_TAG}"
-          else
-            docker buildx imagetools inspect "${IMAGE_NAME}:main"
-          fi

.github/workflows/docs-site-checks.yml

@@ -1,7 +1,13 @@
 name: Docs Site Checks
 
 on:
-  workflow_call:
+  # No paths filter — the job must always run so the required check
+  # reports a status (path-gated workflows leave checks "pending" forever
+  # when no matching files change, which blocks merge).
+  pull_request:
+    branches: [main]
+
+  workflow_dispatch:
 
 permissions:
   contents: read
@@ -9,7 +15,6 @@ permissions:
 jobs:
   docs-site-checks:
     runs-on: ubuntu-latest
-    timeout-minutes: 20
     steps:
       - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
 
@@ -20,19 +25,15 @@ jobs:
           cache-dependency-path: website/package-lock.json
 
       - name: Install website dependencies
-        uses: ./.github/actions/retry
-        with:
-          command: npm ci
-          working-directory: website
+        run: npm ci
+        working-directory: website
 
       - uses: actions/setup-python@a309ff8b426b58ec0e2a45f0f869d46889d02405 # v6.2.0
         with:
           python-version: "3.11"
 
       - name: Install ascii-guard
-        uses: ./.github/actions/retry
-        with:
-          command: python -m pip install ascii-guard==2.3.0 pyyaml==6.0.3
+        run: python -m pip install ascii-guard==2.3.0 pyyaml==6.0.3
 
       - name: Extract skill metadata for dashboard
         run: python3 website/scripts/extract-skills.py

.github/workflows/e2e-desktop.yml

@@ -1,255 +0,0 @@
-name: E2E Desktop
-
-on:
-  workflow_call:
-    outputs:
-      review_status:
-        description: Screenshot and visual-diff status for the CI review comment.
-        value: ${{ jobs.e2e.outputs.review_status }}
-
-permissions:
-  contents: read
-
-concurrency:
-  group: e2e-desktop-${{ github.ref }}
-  cancel-in-progress: true
-
-jobs:
-  e2e:
-    name: Playwright E2E (Linux)
-    runs-on: ubuntu-latest
-    timeout-minutes: 20
-    outputs:
-      review_status: ${{ steps.review-status.outputs.review_status }}
-    steps:
-      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
-
-      # ── System deps for Electron on headless Ubuntu ───────────────────
-      # Electron needs GTK, NSS,atk, etc. even under xvfb. Playwright's
-      # install-deps covers browsers; for Electron we install the apt
-      # packages directly.
-      - name: Install system dependencies for Electron
-        run: |
-          sudo apt-get update -qq
-          sudo apt-get install -y -qq \
-            xvfb \
-            libgtk-3-0 libnotify4 libnss3 libxss1 libxtst6 \
-            xdg-utils libatspi2.0-0 libdrm2 libgbm1 libasound2t64
-
-      # ── Node ───────────────────────────────────────────────────────────
-      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4
-        with:
-          node-version: 22
-          cache: npm
-      # Full npm ci (not --ignore-scripts): electron's postinstall
-      # downloads the binary we launch, and node-pty's native build is
-      # needed for the terminal pane.
-      - uses: ./.github/actions/retry
-        with:
-          command: npm ci
-
-      # ── Python (for the hermes serve backend) ──────────────────────────
-      - name: Install uv
-        uses: astral-sh/setup-uv@fac544c07dec837d0ccb6301d7b5580bf5edae39 # 8.2.0
-        with:
-          enable-cache: true
-          cache-dependency-glob: |
-            pyproject.toml
-            uv.lock
-      - name: Set up Python 3.11
-        run: uv python install 3.11
-      - name: Install Python dependencies
-        uses: ./.github/actions/retry
-        with:
-          command: uv sync --locked --python 3.11 --extra all --extra dev
-
-      # ── Build desktop app ─────────────────────────────────────────────
-      # The Playwright step below runs `npm run build` before testing so
-      # dist/ is always fresh — no separate build step needed here.
-
-      # ── Restore visual baseline screenshots from main ──────────────────
-      # Baselines are generated on main (via --update-snapshots) and cached.
-      # On PRs, we restore them so toHaveScreenshot has something to compare
-      # against.  The cache key is keyed on the desktop source files so a
-      # UI change naturally invalidates it — but we fall back to the main
-      # cache to avoid cold starts on unrelated PRs.
-      - name: Restore visual baseline screenshots
-        id: restore-baselines
-        uses: actions/cache@0400d5f644dc74513175e3cd8d07132dd4860809 # v4.2.4
-        with:
-          path: apps/desktop/e2e/*-snapshots
-          key: visual-baselines-${{ github.ref_name }}
-          restore-keys: |
-            visual-baselines-main
-
-      # ── Run Playwright E2E under xvfb ─────────────────────────────────
-      # xvfb runs at a fixed 1280x1024 screen so the 1220x800 Electron
-      # window always has a consistent viewport for screenshot comparison.
-      # On main, we run with --update-snapshots to generate baselines.
-      # `npm run test:e2e` builds dist/ as a pretest hook so the renderer
-      # is always fresh — no separate build step needed.
-      - name: Run Playwright E2E tests
-        working-directory: apps/desktop
-        run: |
-          if [ "${{ github.ref_name }}" = "main" ]; then
-            echo "On main — generating/updating baseline screenshots"
-            npm run build && xvfb-run -a --server-args="-screen 0 1280x1024x24" \
-              npx playwright test --reporter=list --update-snapshots
-          else
-            echo "On PR — comparing against cached baselines"
-            npm run build && xvfb-run -a --server-args="-screen 0 1280x1024x24" \
-              npx playwright test --reporter=list
-          fi
-        env:
-          CI: "true"
-          # Ensure no real API keys leak into the test env.
-          OPENROUTER_API_KEY: ""
-          OPENAI_API_KEY: ""
-          NOUS_API_KEY: ""
-
-      # ── Save updated baselines to cache (main only) ───────────────────
-      - name: Save updated baselines to cache
-        if: github.ref_name == 'main' && always()
-        uses: actions/cache/save@0400d5f644dc74513175e3cd8d07132dd4860809 # v4.2.4
-        with:
-          path: apps/desktop/e2e/*-snapshots
-          key: visual-baselines-main
-
-      # ── Upload Playwright report (HTML + traces) ──────────────────────
-      - name: Upload Playwright report
-        id: upload-report
-        if: always()
-        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
-        with:
-          name: playwright-report-${{ github.sha }}
-          path: apps/desktop/playwright-report
-          retention-days: 14
-          overwrite: true
-
-      # ── Upload test results (screenshots, traces, diffs) ───────────────
-      - name: Upload test results
-        id: upload-results
-        if: always()
-        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
-        with:
-          name: playwright-test-results-${{ github.sha }}
-          path: apps/desktop/test-results
-          retention-days: 14
-          overwrite: true
-
-      # ── Upload just the visual diffs (small, fast to review) ──────────
-      - name: Upload visual diffs
-        id: upload-diffs
-        if: always() && github.ref_name != 'main'
-        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
-        with:
-          name: visual-diffs-${{ github.sha }}
-          path: |
-            apps/desktop/test-results/**/*-diff.png
-            apps/desktop/test-results/**/*-actual.png
-            apps/desktop/test-results/**/*-expected.png
-          retention-days: 14
-          overwrite: true
-          if-no-files-found: ignore
-
-      - name: Build screenshot review status
-        id: review-status
-        if: always()
-        working-directory: apps/desktop
-        env:
-          RESULTS_URL: ${{ steps.upload-results.outputs.artifact-url }}
-        run: |
-          python3 ../../scripts/ci/e2e_screenshot_status.py \
-            --results-dir test-results \
-            --manifest-output /tmp/e2e-screenshot-manifest.json \
-            --evidence-dir /tmp/e2e-evidence \
-            --artifact-url "$RESULTS_URL" \
-            --output /tmp/e2e-review-status.json
-          {
-            echo 'review_status<<__E2E_REVIEW_STATUS__'
-            cat /tmp/e2e-review-status.json
-            echo '__E2E_REVIEW_STATUS__'
-          } >> "$GITHUB_OUTPUT"
-
-      # The trusted workflow_run publisher consumes only this flat, bounded
-      # artifact. It turns selected images into GitHub attachment URLs; it
-      # never checks out or runs this PR's code.
-      - name: Upload inline E2E evidence
-        if: always() && github.ref_name != 'main'
-        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
-        with:
-          name: e2e-evidence-${{ github.sha }}
-          path: /tmp/e2e-evidence
-          retention-days: 14
-          overwrite: true
-          if-no-files-found: error
-
-      # ── Generate step summary with visual diff info ───────────────────
-      # Parse the JSON report + scan for diff images, then post a summary
-      # to the GitHub Actions step output so reviewers can see what changed
-      # without downloading artifacts.  Runs AFTER uploads so it can link
-      # the artifact download URLs from their step outputs.
-      - name: Generate visual diff summary
-        if: always()
-        working-directory: apps/desktop
-        env:
-          REPORT_URL: ${{ steps.upload-report.outputs.artifact-url }}
-          RESULTS_URL: ${{ steps.upload-results.outputs.artifact-url }}
-          DIFFS_URL: ${{ steps.upload-diffs.outputs.artifact-url }}
-        run: |
-          {
-          echo "## Desktop E2E — Visual Diff Report"
-          echo ""
-
-          # Count diff images (playwright writes *-diff.png on mismatch)
-          DIFF_COUNT=$(find test-results -name '*-diff.png' 2>/dev/null | wc -l)
-          ACTUAL_COUNT=$(find test-results -name '*-actual.png' 2>/dev/null | wc -l)
-
-          if [ "$DIFF_COUNT" -eq 0 ]; then
-            echo "✅ All $ACTUAL_COUNT screenshot(s) matched their baselines (or no baselines existed yet)."
-          else
-            echo "📸 **$DIFF_COUNT of $ACTUAL_COUNT screenshot(s) differ from baseline:**"
-            echo ""
-            echo "| Test | Diff | Actual | Expected |"
-            echo "|------|------|--------|----------|"
-
-            # List each diff image with a link to the artifact
-            for diff in $(find test-results -name '*-diff.png' 2>/dev/null | sort); do
-              base=${diff%-diff.png}
-              test_name=$(basename "$base")
-              echo "| $test_name | [diff]($diff) | [actual](${base}-actual.png) | [expected](${base}-expected.png) |"
-            done
-          fi
-
-          echo ""
-          echo "📥 **Artifacts:**"
-          echo ""
-          if [ -n "$RESULTS_URL" ]; then
-            echo "- [playwright-test-results]($RESULTS_URL) — all screenshots (actual + expected + diff) + traces"
-          fi
-          if [ -n "$REPORT_URL" ]; then
-            echo "- [playwright-report]($REPORT_URL) — interactive HTML report"
-          fi
-          if [ -n "$DIFFS_URL" ]; then
-            echo "- [visual-diffs]($DIFFS_URL) — just the diffed screenshots (small, fast to review)"
-          fi
-          echo ""
-          echo "**To update baselines:** merge to main (baselines auto-update on main runs) or run \`npx playwright test --update-snapshots\` locally."
-
-          # Also parse the JSON report for pass/fail counts
-          if [ -f playwright-report/results.json ]; then
-            echo ""
-            echo "### Test Results"
-            echo ""
-            node -e "
-              const r = require('./playwright-report/results.json');
-              const stats = r.stats || {};
-              console.log('| Status | Count |');
-              console.log('|--------|-------|');
-              console.log('| ✅ Passed | ' + (stats.expected || 0) + ' |');
-              console.log('| ❌ Failed | ' + (stats.unexpected || 0) + ' |');
-              console.log('| ⏭️ Skipped | ' + (stats.skipped || 0) + ' |');
-              console.log('| 🔄 Flaky | ' + (stats.flaky || 0) + ' |');
-            " 2>/dev/null || true
-          fi
-          } >> "$GITHUB_STEP_SUMMARY"

.github/workflows/history-check.yml

@@ -14,11 +14,11 @@ name: History Check
 # the PR head and main to be non-empty.
 
 on:
-  workflow_call:
-    outputs:
-      review_status:
-        description: "JSON array of review_status objects for the synthesizer."
-        value: ${{ jobs.check-common-ancestor.outputs.review_status }}
+  # No paths filter — the job must always run so the required check
+  # reports a status (path-gated workflows leave checks "pending" forever
+  # when no matching files change, which blocks merge).
+  pull_request:
+    branches: [main]
 
 permissions:
   contents: read
@@ -26,24 +26,18 @@ permissions:
 jobs:
   check-common-ancestor:
     runs-on: ubuntu-latest
-    timeout-minutes: 10
-    outputs:
-      review_status: ${{ steps.merge-base-check.outputs.review_status }}
     steps:
       - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
         with:
           fetch-depth: 0 # full history both sides for merge-base
 
-      - id: merge-base-check
-        name: Reject PRs with no common ancestor on main
+      - name: Reject PRs with no common ancestor on main
         run: |
           # `git merge-base` exits non-zero AND prints nothing when the two
           # commits share no ancestor.  We check both conditions explicitly
           # so the failure message is clear regardless of which signal fires
           # first.
           if ! BASE=$(git merge-base origin/main HEAD 2>/dev/null) || [ -z "$BASE" ]; then
-            STATUS='[{"source":"unrelated histories","results":[{"kind":"action_required","title":"Unrelated histories","summary":"This PR has no common ancestor with main.","detail":"","how_to_fix":"Rebase your changes onto current main:\n```\ngit fetch origin main\ngit checkout -b fix-branch origin/main\n# re-apply your changes (cherry-pick, copy files, etc.)\ngit push -f origin fix-branch\n```\n"}]}]'
-            echo "review_status=${STATUS}" >> "$GITHUB_OUTPUT"
             echo ""
             echo "::error::This PR has no common ancestor with main."
             echo ""
@@ -65,4 +59,3 @@ jobs:
             exit 1
           fi
           echo "::notice::Common ancestor with main: $BASE"
-          echo "review_status=[]" >> "$GITHUB_OUTPUT"

.github/workflows/js-autofix.yml

@@ -1,259 +0,0 @@
-name: auto-fix lint issues & formatting
-
-# On push to main (or manual trigger), run `npm run fix` on each workspace
-# package and apply any changes via a PR.
-#
-# Fixable lint issues (import sorting, unused imports, curly braces, etc.) are
-# auto-corrected on merge so PRs aren't blocked by them. The PR-time eslint
-# check in typecheck.yml fails only when un-fixable errors remain.
-#
-# NOTE: App token pushes DO trigger further workflow runs (unlike
-# secrets.GITHUB_TOKEN). The concurrency group (ts-autofix-${{ github.ref }})
-# with cancel-in-progress: true prevents an infinite loop — a re-triggered
-# run cancels the in-flight one, and since the second run finds no new fixes
-# (the first run already applied them), it exits with an empty patch.
-#
-# ── Security model: two-job split ───────────────────────────────────────────
-#
-# The eslint process executes repo code (eslint.config.mjs, package.json
-# scripts, installed plugins). To prevent a malicious PR from getting arbitrary
-# code execution on a runner with push access, the work is split:
-#
-#   1. generate-patch (unprivileged, contents: read only)
-#      Checks out, installs deps, runs eslint --fix, produces a .patch artifact.
-#      Worst case: malicious code runs here on an ephemeral runner with zero
-#      push permissions.
-#
-#   2. apply-patch (privileged, contents: write + pull-requests: write)
-#      Checks out, downloads the patch artifact, applies it, pushes to the
-#      bot/js-autofix branch, creates/updates a PR, and enables auto-merge.
-#      This job never runs npm, never installs anything, never executes any
-#      repo code. The only input it trusts is the patch artifact.
-#      Skipped entirely when generate-patch reports no fixes (has-fixes != true).
-#      The PR auto-merges (squash) once CI passes. If CI fails or main moves,
-#      the PR is auto-closed and the branch deleted — the next run re-applies
-#      on the current state.
-
-on:
-  push:
-    branches: [main]
-    paths:
-      - '**/*.js'
-      - '**/*.cjs'
-      - '**/*.mjs'
-      - '**/*.ts'
-      - '**/*.tsx'
-      - 'package.json'
-      - 'package-lock.json'
-  workflow_dispatch:
-
-permissions:
-  contents: read # default; apply-patch job overrides to write
-
-concurrency:
-  group: ts-autofix-${{ github.ref }}
-  cancel-in-progress: true
-
-jobs:
-  generate-patch:
-    name: Generate eslint --fix patch
-    runs-on: ubuntu-latest
-    timeout-minutes: 5
-    outputs:
-      has-fixes: ${{ steps.produce-patch.outputs.has-fixes }}
-    # No permissions override → inherits workflow-level contents: read.
-    steps:
-      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
-
-      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4
-        with:
-          node-version: 22
-          cache: npm
-
-      # --ignore-scripts: eslint only needs TS sources + eslint packages.
-      - uses: ./.github/actions/retry
-        with:
-          command: npm ci --ignore-scripts
-
-      - name: npm run fix in all workspaces
-        # continue-on-error: if un-fixable errors exist on main, we still want
-        # to commit whatever fixes were applied. The PR-time check in
-        # typecheck.yml is what blocks un-fixable errors from landing.
-        continue-on-error: true
-        run: npm run fix
-
-      - name: Produce patch
-        id: produce-patch
-        run: |
-          if git diff --quiet; then
-            echo "No fixes needed."
-            echo "has-fixes=false" >> "$GITHUB_OUTPUT"
-            # Empty patch signals "nothing to do" to apply-patch.
-            : > js-fix.patch
-          else
-            git diff > js-fix.patch
-            echo "has-fixes=true" >> "$GITHUB_OUTPUT"
-            echo "Patch size: $(wc -c < js-fix.patch) bytes"
-
-            # Reject patches that touch anything outside JS/TS/JSON sources.
-            # `npm run fix` should only ever modify those; anything else means
-            # eslint/prettier or a plugin went rogue and we refuse to ship it.
-            BAD=$(git diff --name-only | grep -vE '\.(js|cjs|mjs|ts|tsx|json)$' || true)
-            if [ -n "$BAD" ]; then
-              echo "::error::Refusing to upload patch — touches disallowed files:"
-              echo "$BAD"
-              exit 1
-            fi
-          fi
-
-      - name: Upload patch artifact
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4
-        with:
-          name: js-fix-patch
-          path: js-fix.patch
-          retention-days: 1
-          include-hidden-files: true
-
-  apply-patch:
-    name: Apply patch
-    needs: generate-patch
-    # Skip entirely when generate-patch found no fixes — saves a runner,
-    # avoids a redundant checkout/download, and keeps the job graph honest.
-    if: needs.generate-patch.outputs.has-fixes == 'true'
-    runs-on: ubuntu-latest
-    timeout-minutes: 15
-    environment: trusted-automation
-    permissions:
-      contents: write # needed to push to bot/js-autofix
-      pull-requests: write # needed for PR creation + auto-merge
-    steps:
-      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
-
-      - name: Get GitHub App token
-        id: app-token
-        uses: ./.github/actions/get-app-token
-        with:
-          client-id: ${{ vars.APP_CLIENT_ID }}
-          private-key: ${{ secrets.APP_PRIVATE_KEY }}
-
-      - name: Download patch
-        uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4
-        with:
-          name: js-fix-patch
-          # ${{ runner.temp }} expands in with: params (shell-style $VAR does not).
-          # download-artifact's path is a *directory* — the artifact's js-fix.patch
-          # file lands inside it, so $RUNNER_TEMP/js-fix.patch resolves correctly
-          # in the run step below.
-          path: ${{ runner.temp }}
-
-      - name: Apply patch and push to bot branch
-        env:
-          BOT_BRANCH: bot/js-autofix
-        run: |
-          set -euo pipefail
-
-          # Empty patch = nothing to do.
-          if [ ! -s "$RUNNER_TEMP/js-fix.patch" ]; then
-            echo "Patch is empty. No fixes to apply."
-            exit 0
-          fi
-
-          # Apply the patch produced by the unprivileged job.
-          git apply --check "$RUNNER_TEMP/js-fix.patch" || {
-            echo "::error::Patch does not apply cleanly. Branch may have moved."
-            exit 1
-          }
-          git apply "$RUNNER_TEMP/js-fix.patch"
-
-          git config user.name "github-actions[bot]"
-          git config user.email "github-actions[bot]@users.noreply.github.com"
-          git add -A
-          git commit -m "fmt(js): \`npm run fix\` on merge"
-
-          # Push to the dedicated bot branch. Force-push is safe here:
-          # bot/js-autofix is a bot-only branch that gets rewritten each run.
-          # If the branch was deleted after a previous PR merge, this
-          # recreates it.
-          git push --force origin HEAD:"$BOT_BRANCH"
-
-      - name: Create/update PR and enable auto-merge
-        env:
-          GH_TOKEN: ${{ steps.app-token.outputs.token }}
-          BOT_BRANCH: bot/js-autofix
-        run: |
-          set -euo pipefail
-
-          # Create PR if one doesn't exist. If it already exists, the
-          # force-push above already updated it with the latest fixes.
-          PR_NUM=$(gh pr list --head "$BOT_BRANCH" --state open --json number --jq '.[0].number' 2>/dev/null || true)
-          if [ -z "$PR_NUM" ]; then
-            # gh pr create prints the PR URL. Extract the number from it
-            # (https://github.com/<org>/<repo>/pull/<number>).
-            PR_URL=$(gh pr create \
-              --head "$BOT_BRANCH" --base main \
-              --title 'fmt(js): `npm run fix` auto-fix' \
-              --body 'Auto-generated by the `auto-fix lint issues & formatting` workflow. Auto-merges (squash) once CI passes. If CI fails or `main` moves, the PR is auto-closed and the branch deleted — the next run re-applies on the current state.')
-            PR_NUM=$(echo "$PR_URL" | grep -oE '[0-9]+$')
-          fi
-
-          # Enable auto-merge (squash). If already enabled, this is a no-op.
-          gh pr merge "$PR_NUM" --auto --squash || true
-
-      - name: Wait for merge, auto-close on failure or stale
-        env:
-          GH_TOKEN: ${{ steps.app-token.outputs.token }}
-          START_SHA: ${{ github.sha }}
-        run: |
-          set -euo pipefail
-
-          PR_NUM=$(gh pr list --head bot/js-autofix --state open --json number --jq '.[0].number' 2>/dev/null || true)
-          if [ -z "$PR_NUM" ]; then
-            echo "No open PR. Nothing to wait for."
-            exit 0
-          fi
-
-          echo "Waiting for PR #$PR_NUM to merge..."
-
-          # Poll every 15s for up to ~10 minutes. Auto-merge will handle the
-          # PR even if this job times out — the polling is for cleanup only
-          # (auto-close on CI failure, conflicts, or main moving).
-          for i in $(seq 1 40); do
-            sleep 15
-
-            STATE=$(gh pr view "$PR_NUM" --json state --jq '.state')
-            if [ "$STATE" = "MERGED" ] || [ "$STATE" = "CLOSED" ]; then
-              echo "PR #$PR_NUM is $STATE."
-              exit 0
-            fi
-
-            # If main moved, the PR may have already merged (which moves
-            # main) or another commit landed. Re-check state first.
-            CURRENT_SHA=$(gh api "repos/${{ github.repository }}/branches/main" --jq '.commit.sha')
-            if [ "$CURRENT_SHA" != "$START_SHA" ]; then
-              STATE=$(gh pr view "$PR_NUM" --json state --jq '.state')
-              if [ "$STATE" = "MERGED" ]; then
-                echo "PR #$PR_NUM merged (main moved to $CURRENT_SHA)."
-                exit 0
-              fi
-              echo "Main moved ($START_SHA → $CURRENT_SHA). Closing stale PR."
-              gh pr close "$PR_NUM" --delete-branch || true
-              exit 0
-            fi
-
-            # If CI checks failed, close + delete the branch.
-            if gh pr checks "$PR_NUM" 2>/dev/null | grep -qi "fail"; then
-              echo "CI failed on PR #$PR_NUM. Closing + deleting branch."
-              gh pr close "$PR_NUM" --delete-branch
-              exit 0
-            fi
-
-            # If PR is conflicted, close + delete the branch.
-            MERGEABLE=$(gh pr view "$PR_NUM" --json mergeable --jq '.mergeable')
-            if [ "$MERGEABLE" = "CONFLICTING" ]; then
-              echo "PR #$PR_NUM is conflicted. Closing + deleting branch."
-              gh pr close "$PR_NUM" --delete-branch
-              exit 0
-            fi
-          done
-
-          echo "Timeout reached. Auto-merge will handle PR #$PR_NUM if CI passes."

.github/workflows/js-tests.yml

@@ -1,69 +0,0 @@
-# .github/workflows/js-tests.yml
-name: JS Tests
-
-on:
-  workflow_call:
-
-jobs:
-  workspaces:
-    name: List npm workspaces
-    runs-on: ubuntu-latest
-    timeout-minutes: 20
-    outputs:
-      checks: ${{ steps.set-matrix.outputs.checks }}
-    steps:
-      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
-      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4
-        with:
-          node-version: 22
-          cache: npm
-      - uses: ./.github/actions/retry
-        with:
-          command: npm ci --ignore-scripts
-      - id: set-matrix
-        run: |
-          node -e '
-          const { execSync } = require("child_process");
-          const pkgs = JSON.parse(execSync("npm query .workspace", { encoding: "utf-8" }));
-          if (pkgs.length === 0) {
-            console.error("::error::Workspace discovery produced an empty package list — refusing to emit a zero-length matrix (would skip all JS/TS checks silently).");
-            process.exit(1);
-          }
-          const checks = [];
-          for (const pkg of pkgs) {
-            const scripts = pkg.scripts || {};
-            const subs = Object.keys(scripts).filter(s => /^check:.+$/.test(s));
-            if (subs.length > 0) {
-              for (const script of subs) {
-                checks.push({ package: pkg.location, script });
-              }
-            } else if (scripts.check) {
-              checks.push({ package: pkg.location, script: "check" });
-            }
-          }
-          if (checks.length === 0) {
-            console.error("::error::No check scripts found in any workspace package.");
-            process.exit(1);
-          }
-          process.stdout.write("checks=" + JSON.stringify(checks) + "\n");
-          ' >> "$GITHUB_OUTPUT"
-
-  check:
-    name: ${{ matrix.package }} / ${{ matrix.script }}
-    needs: workspaces
-    runs-on: ubuntu-latest
-    timeout-minutes: 20
-    strategy:
-      matrix:
-        include: ${{ fromJson(needs.workspaces.outputs.checks) }}
-      fail-fast: false # report all failures, not just the first one
-    steps:
-      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
-      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4
-        with:
-          node-version: 22
-          cache: npm
-      - uses: ./.github/actions/retry
-        with:
-          command: npm ci
-      - run: npm run --prefix ${{ matrix.package }} ${{ matrix.script }}

.github/workflows/label-rerun.yml

@@ -1,81 +0,0 @@
-name: Label rerun
-
-# When the ``ci-reviewed`` label is added to a PR, rerun all failed jobs in
-# the latest CI run. This re-evaluates ``review-labels`` (which now sees the
-# label) and GitHub automatically reruns dependent jobs (``comment-live``,
-# ``all-checks-pass``) — so the review comment gets updated too.
-#
-# If the CI run is still in progress when the label is added, we wait for it
-# to finish before rerunning (``gh run rerun`` only works on completed runs).
-# The wait can be long (20+ min for a full CI run), but it's better than
-# silently failing and leaving the reviewer stuck.
-
-on:
-  pull_request:
-    types: [labeled]
-
-permissions:
-  actions: write
-  pull-requests: read
-
-concurrency:
-  group: label-rerun-${{ github.event.pull_request.number }}
-  cancel-in-progress: true
-
-jobs:
-  rerun-review-labels:
-    name: Rerun review-labels job
-    if: github.event.label.name == 'ci-reviewed'
-    runs-on: ubuntu-latest
-    timeout-minutes: 40
-    steps:
-      - name: Wait for CI run to finish, then rerun failed jobs
-        env:
-          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-          REPO: ${{ github.repository }}
-          HEAD_SHA: ${{ github.event.pull_request.head.sha }}
-        run: |
-          set -uo pipefail
-
-          # Find the latest CI run for this PR's head SHA.
-          RUN_ID=$(gh run list \
-            --repo "$REPO" \
-            --commit "$HEAD_SHA" \
-            --workflow ci.yml \
-            --limit 1 \
-            --json databaseId,status \
-            --jq '.[0] | "\(.databaseId) \(.status)"' 2>/dev/null || true)
-
-          if [ -z "$RUN_ID" ]; then
-            echo "No CI run found for this PR — nothing to rerun."
-            exit 0
-          fi
-
-          # Split "RUN_ID STATUS" into two vars.
-          RUN_ID="${RUN_ID%% *}"
-          STATUS="${RUN_ID##* }"
-
-          echo "Latest CI run: $RUN_ID (status: $STATUS)"
-
-          # If the run is still in progress, wait for it to finish.
-          # gh run rerun only works on completed runs — if we try while it's
-          # running, GitHub rejects with "cannot be rerun; This workflow is
-          # already running".
-          if [ "$STATUS" != "completed" ]; then
-            echo "Run is $STATUS — waiting for completion (this may take a while)..."
-            # gh run watch --exit-status exits non-zero if the run fails,
-            # which is expected (the label gate fails). Don't let that kill
-            # the workflow — we WANT to rerun failed jobs.
-            timeout 2100 gh run watch "$RUN_ID" --repo "$REPO" --interval 15 || true
-
-            # Verify it's actually completed now.
-            STATUS=$(gh run view "$RUN_ID" --repo "$REPO" --json status --jq '.status' 2>/dev/null || echo "unknown")
-            if [ "$STATUS" != "completed" ]; then
-              echo "Run is still $STATUS after wait — giving up."
-              exit 0
-            fi
-          fi
-
-          echo "Run completed. Rerunning all failed jobs..."
-          gh run rerun "$RUN_ID" --repo "$REPO" --failed || true
-          echo "Done. GitHub will rerun review-labels and all dependent jobs."

.github/workflows/lint.yml

@@ -2,25 +2,29 @@ name: Lint (ruff + ty)
 
 # Two things here:
 #   1. Advisory diff — ruff + ty diagnostics as a diff vs the target branch.
-#      Writes a Markdown summary to the run page. Exit zero always.
+#      Posts a Markdown summary and a PR comment. Exit zero always.
 #   2. Blocking ``ruff check .`` — enforces the explicit rules in
 #      ``[tool.ruff.lint.select]`` (currently PLW1514). Failure blocks merge.
-#      Separate job so the advisory diff still runs even when enforcement
-#      fails.
-#
-# CI-sensitive file review was previously here as a ``ci-review`` job but
-# has moved to ``review-labels.yml`` so it can be rerun independently.
+#      Separate job so the advisory diff still runs and posts even when
+#      enforcement fails.
 
 on:
-  workflow_call:
-    inputs:
-      event_name:
-        description: The event name from the calling orchestrator (pull_request or push).
-        type: string
-        required: true
+  push:
+    branches: [main]
+    paths-ignore:
+      - "**/*.md"
+      - "docs/**"
+      - "website/**"
+
+  # No paths filter — the job must always run so the required check
+  # reports a status (path-gated workflows leave checks "pending" forever
+  # when no matching files change, which blocks merge).
+  pull_request:
+    branches: [main]
 
 permissions:
   contents: read
+  pull-requests: write # needed to post/update PR comments
 
 concurrency:
   group: lint-${{ github.ref }}
@@ -29,7 +33,6 @@ concurrency:
 jobs:
   lint-diff:
     name: ruff + ty diff
-    if: inputs.event_name == 'pull_request'
     runs-on: ubuntu-latest
     timeout-minutes: 10
     steps:
@@ -39,19 +42,19 @@ jobs:
           fetch-depth: 0 # need full history for merge-base + worktree
 
       - name: Install uv
-        uses: astral-sh/setup-uv@fac544c07dec837d0ccb6301d7b5580bf5edae39 # 8.2.0
+        uses: astral-sh/setup-uv@d4b2f3b6ecc6e67c4457f6d3e41ec42d3d0fcb86 # v5
 
       - name: Install ruff + ty
-        uses: ./.github/actions/retry
-        with:
-          command: uv tool install ruff && uv tool install ty
+        run: |
+          uv tool install ruff
+          uv tool install ty
 
       - name: Determine base ref
         id: base
         run: |
           # For PRs, diff against the merge base with the target branch.
           # For pushes to main, diff against the previous commit on main.
-          if [ "${{ inputs.event_name }}" = "pull_request" ]; then
+          if [ "${{ github.event_name }}" = "pull_request" ]; then
             BASE_SHA=$(git merge-base "origin/${{ github.base_ref }}" HEAD)
             BASE_REF="origin/${{ github.base_ref }}"
           else
@@ -100,8 +103,6 @@ jobs:
           echo "base ty:   $(wc -c < .lint-reports/base/ty.json) bytes"
 
       - name: Generate diff summary
-        env:
-          HEAD_REF: ${{ inputs.event_name == 'pull_request' && github.head_ref || github.ref_name }}
         run: |
           python scripts/lint_diff.py \
             --base-ruff .lint-reports/base/ruff.json \
@@ -109,10 +110,50 @@ jobs:
             --base-ty   .lint-reports/base/ty.json \
             --head-ty   .lint-reports/head/ty.json \
             --base-ref  "${{ steps.base.outputs.ref }}" \
-            --head-ref  "$HEAD_REF" \
+            --head-ref  "${{ github.event_name == 'pull_request' && github.head_ref || github.ref_name }}" \
             --output    .lint-reports/summary.md
           cat .lint-reports/summary.md >> "$GITHUB_STEP_SUMMARY"
 
+      - name: Upload reports as artifact
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4
+        with:
+          name: lint-reports
+          path: .lint-reports/
+          retention-days: 14
+
+      - name: Post / update PR comment
+        if: github.event_name == 'pull_request' && github.event.pull_request.head.repo.full_name == github.repository
+        continue-on-error: true
+        uses: actions/github-script@60a0d83039c74a4aee543508d2ffcb1c3799cdea # v7
+        with:
+          script: |
+            const fs = require('fs');
+            const body = fs.readFileSync('.lint-reports/summary.md', 'utf8');
+            const marker = '<!-- lint-diff-summary -->';
+            const fullBody = marker + '\n' + body;
+
+            const { data: comments } = await github.rest.issues.listComments({
+              owner: context.repo.owner,
+              repo:  context.repo.repo,
+              issue_number: context.issue.number,
+            });
+            const existing = comments.find(c => c.body && c.body.includes(marker));
+            if (existing) {
+              await github.rest.issues.updateComment({
+                owner: context.repo.owner,
+                repo:  context.repo.repo,
+                comment_id: existing.id,
+                body: fullBody,
+              });
+            } else {
+              await github.rest.issues.createComment({
+                owner: context.repo.owner,
+                repo:  context.repo.repo,
+                issue_number: context.issue.number,
+                body: fullBody,
+              });
+            }
+
   ruff-blocking:
     # Enforce the rules in pyproject.toml [tool.ruff.lint.select]. Currently
     # PLW1514 (unspecified-encoding) — catches bare ``open()`` /
@@ -128,12 +169,10 @@ jobs:
         uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
 
       - name: Install uv
-        uses: astral-sh/setup-uv@fac544c07dec837d0ccb6301d7b5580bf5edae39 # 8.2.0
+        uses: astral-sh/setup-uv@d4b2f3b6ecc6e67c4457f6d3e41ec42d3d0fcb86 # v5
 
       - name: Install ruff
-        uses: ./.github/actions/retry
-        with:
-          command: uv tool install ruff
+        run: uv tool install ruff
 
       - name: ruff check .
         # No --exit-zero, no || true. Exit code propagates to the job,

.github/workflows/lockfile-diff.yml

@@ -1,95 +0,0 @@
-name: Lockfile diff
-
-# Advisory PR comment showing the *semantic* diff of package-lock.json
-# changes — which packages were added/removed/updated and their versions.
-# The raw textual diff of a lockfile is unreadable (npm reorders entries
-# and rewrites integrity hashes), so scripts/ci/lockfile_diff.py parses
-# the ``packages`` map at the merge base and at HEAD and set-diffs the
-# {install path: version} maps instead.
-#
-# The semantic diff is exposed as a workflow_call output ``review_status``
-# (a JSON array in the unified status format) and an artifact
-# (``lockfile-diff`` containing the markdown fragment) for the step
-# summary.
-#
-# Never blocking — this is review signal, not enforcement.
-
-on:
-  workflow_call:
-    outputs:
-      changed:
-        description: Whether package-lock.json changed relative to the target branch.
-        value: ${{ jobs.diff.outputs.changed }}
-      review_status:
-        description: JSON array of review status objects for the unified PR comment.
-        value: ${{ jobs.diff.outputs.review_status }}
-
-permissions:
-  contents: read
-
-concurrency:
-  group: lockfile-diff-${{ github.event.pull_request.number || github.ref }}
-  cancel-in-progress: true
-
-jobs:
-  diff:
-    name: package-lock.json semantic diff
-    runs-on: ubuntu-latest
-    timeout-minutes: 5
-    outputs:
-      changed: ${{ steps.diff.outputs.changed }}
-      review_status: ${{ steps.emit-status.outputs.review_status }}
-    steps:
-      - name: Checkout code
-        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
-        with:
-          fetch-depth: 0 # need history for the merge base
-
-      - name: Generate semantic lockfile diff
-        id: diff
-        run: |
-          set -euo pipefail
-          # Three-dot semantics by hand: diff from the merge base with the
-          # target branch to the PR head, so changes that landed on main
-          # after the branch point don't show up as this PR's doing.
-          BASE_SHA=$(git merge-base "origin/${{ github.base_ref }}" HEAD)
-          echo "Merge base: ${BASE_SHA}"
-          python3 scripts/ci/lockfile_diff.py \
-            --base "$BASE_SHA" \
-            --head HEAD \
-            --output /tmp/lockfile-diff.md
-          if [ -s /tmp/lockfile-diff.md ]; then
-            echo "changed=true" >> "$GITHUB_OUTPUT"
-            {
-              echo "## package-lock.json semantic diff"
-              echo ""
-              cat /tmp/lockfile-diff.md
-            } >> "$GITHUB_STEP_SUMMARY"
-          else
-            echo "changed=false" >> "$GITHUB_OUTPUT"
-            : > /tmp/lockfile-diff.md
-          fi
-
-      - name: Emit review_status
-        id: emit-status
-        run: |
-          set -euo pipefail
-          CHANGED="${{ steps.diff.outputs.changed }}"
-          STATUS="[]"
-
-          if [ "$CHANGED" = "true" ]; then
-            CONTENT=$(cat /tmp/lockfile-diff.md | python3 -c "import sys,json; print(json.dumps(sys.stdin.read()))")
-            STATUS="[{\"source\":\"lockfile-diff\",\"results\":[{\"kind\":\"action_required\",\"title\":\"package-lock.json\",\"summary\":\"Locked npm dependency versions changed.\",\"detail\":${CONTENT},\"how_to_fix\":\"Add the \`ci-reviewed\` label after verifying the version changes are expected.\"}]}"
-          else
-            STATUS="[]"
-          fi
-
-          echo "review_status=${STATUS}" >> "$GITHUB_OUTPUT"
-
-      - name: Upload diff artifact
-        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
-        with:
-          name: lockfile-diff
-          path: /tmp/lockfile-diff.md
-          retention-days: 1
-          overwrite: true

.github/workflows/osv-scanner.yml

@@ -1,8 +1,8 @@
 name: OSV-Scanner
 
 # Scans lockfiles (uv.lock, package-lock.json) against the OSV vulnerability
-# database. Runs on every PR/push (via the ci.yml orchestrator's workflow_call)
-# and on a weekly schedule against main.
+# database. Runs on every PR that touches a lockfile and on a weekly schedule
+# against main.
 #
 # This is detection-only — OSV-Scanner does NOT open PRs or modify pins.
 # It reports known CVEs in currently-pinned dependency versions so we can
@@ -10,29 +10,37 @@ name: OSV-Scanner
 # (full SHA / exact version) is preserved; only the notification signal
 # is added.
 #
-# Complements the supply-chain-audit.yml workflow (which scans for malicious
-# code patterns in PR diffs) by covering the orthogonal "currently-pinned
-# dep became known-vulnerable" case.
+# Complements the existing supply-chain-audit.yml workflow (which scans
+# for malicious code patterns in PR diffs) by covering the orthogonal
+# "currently-pinned dep became known-vulnerable" case.
 #
 # Uses Google's officially-recommended reusable workflow, pinned by SHA.
 # Findings land in the repo's Security tab (Code Scanning > OSV-Scanner).
 # fail-on-vuln is disabled so the job does not block merges on pre-existing
 # vulnerabilities in pinned deps that we may need to patch deliberately.
-#
-# The reusable workflow can't emit custom outputs, so a wrapper job
-# downloads the SARIF result and summarizes the vulnerability count into
-# a review_status for the unified PR comment.
 
 on:
-  workflow_call:
+  # No paths filter — the job must always run so the required check
+  # reports a status (path-gated workflows leave checks "pending" forever
+  # when no matching files change, which blocks merge).
+  pull_request:
+    branches: [main]
+  push:
+    branches: [main]
+    paths:
+      - "uv.lock"
+      - "pyproject.toml"
+      - "package.json"
+      - "package-lock.json"
+      - "website/package-lock.json"
   schedule:
     # Weekly scan against main — catches CVEs published after merge for
     # deps that haven't changed since.
-    - cron: '0 9 * * 1'
+    - cron: "0 9 * * 1"
   workflow_dispatch:
 
 permissions:
-  # Required to upload SARIF file to CodeQL. See: https://github.com/github/codeql-action/issues/2117
+  # Required by the reusable workflow to upload SARIF to the Security tab.
   actions: read
   contents: read
   security-events: write
@@ -48,80 +56,4 @@ jobs:
         --lockfile=uv.lock
         --lockfile=package-lock.json
         --lockfile=website/package-lock.json
-      # The upstream reusable workflow uploads this exact file under its
-      # fixed artifact name, which the wrapper downloads below.
-      results-file-name: osv-results.sarif
       fail-on-vuln: false
-
-  emit-status:
-    name: Emit review status
-    runs-on: ubuntu-latest
-    needs: scan
-    if: always()
-    outputs:
-      review_status: ${{ steps.emit.outputs.review_status }}
-    steps:
-      - name: Checkout code
-        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
-
-      - name: Download SARIF result
-        uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4
-        with:
-          name: OSV Scanner SARIF file
-          path: /tmp/osv-results
-        continue-on-error: true
-
-      - name: Emit review_status
-        id: emit
-        run: |
-          set -euo pipefail
-          STATUS="[]"
-
-          if [ -f /tmp/osv-results/osv-results.sarif ]; then
-            # Count vulnerabilities from the SARIF file
-            VULN_COUNT=$(python3 -c "
-          import json, sys
-          try:
-              with open('/tmp/osv-results/osv-results.sarif') as f:
-                  data = json.load(f)
-              count = 0
-              vulns = []
-              for run in data.get('runs', []):
-                  for result in run.get('results', []):
-                      count += 1
-                      rule_id = result.get('ruleId', 'unknown')
-                      message = result.get('message', {}).get('text', '')
-                      loc = result.get('locations', [{}])[0].get('physicalLocation', {}).get('artifactLocation', {}).get('uri', '')
-                      vulns.append(f'- {rule_id} in {loc}: {message}')
-              print(count)
-              if vulns:
-                  print('\n'.join(vulns[:20]), file=sys.stderr)
-          except Exception:
-              print(0)
-          ")
-
-            VULN_DETAIL=""
-            if [ "$VULN_COUNT" -gt 0 ] 2>/dev/null; then
-              VULN_PLURAL=$([ "$VULN_COUNT" -eq 1 ] && echo "y" || echo "ies")
-              VULN_DETAIL=$(python3 -c "
-          import json, sys
-          try:
-              with open('/tmp/osv-results/osv-results.sarif') as f:
-                  data = json.load(f)
-              vulns = []
-              for run in data.get('runs', []):
-                  for result in run.get('results', []):
-                      rule_id = result.get('ruleId', 'unknown')
-                      loc = result.get('locations', [{}])[0].get('physicalLocation', {}).get('artifactLocation', {}).get('uri', '')
-                      vulns.append(f'- {rule_id} in {loc}')
-              print(json.dumps('\n'.join(vulns[:20])))
-          except Exception:
-              print(json.dumps(''))
-          ")
-              STATUS="[{\"source\":\"osv scan\",\"results\":[{\"kind\":\"warning\",\"title\":\"OSV vulnerability scan\",\"summary\":\"${VULN_COUNT} known vulnerabilit${VULN_PLURAL} found in pinned dependencies.\",\"detail\":${VULN_DETAIL},\"how_to_fix\":\"Review the findings in the [Security tab](../../security/code-scanning). Update the affected dependencies if a patched version is available.\"}]}]"
-            else
-              STATUS="[]"
-            fi
-          fi
-
-          echo "review_status=${STATUS}" >> "$GITHUB_OUTPUT"

.github/workflows/publish-e2e-evidence.yml

@@ -1,71 +0,0 @@
-name: Publish E2E evidence
-
-# This runs only from the default branch after CI completes. It intentionally
-# checks out main, never the PR ref, and treats the downloaded artifact as
-# untrusted input before uploading validated GitHub attachments.
-on:
-  workflow_run:
-    workflows: [CI]
-    types: [completed]
-
-permissions:
-  actions: read
-  contents: read
-  pull-requests: write
-
-concurrency:
-  group: publish-e2e-evidence-${{ github.event.workflow_run.id }}
-  cancel-in-progress: false
-
-jobs:
-  publish:
-    name: Publish inline E2E evidence
-    if: github.event.workflow_run.event == 'pull_request'
-    runs-on: ubuntu-latest
-    timeout-minutes: 10
-    environment: gh-image
-    steps:
-      - name: Check out trusted publisher
-        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
-        with:
-          ref: ${{ github.event.repository.default_branch }}
-          persist-credentials: false
-
-      # v1.2.0 resolves to 44f4b93ecbbe22de6c45fa2f62f519aee564ca8c.
-      - name: Install gh-image
-        env:
-          GH_TOKEN: ${{ github.token }}
-        run: gh extension install drogers0/gh-image --pin v1.2.0
-
-      - name: Download and attach evidence
-        env:
-          GH_TOKEN: ${{ github.token }}
-          GITHUB_TOKEN: ${{ github.token }}
-          GH_SESSION_TOKEN: ${{ secrets.GH_IMAGE_SESSION_TOKEN }}
-          SOURCE_REPO: ${{ github.repository }}
-          SOURCE_RUN_ID: ${{ github.event.workflow_run.id }}
-        run: |
-          set -euo pipefail
-
-          PR_NUMBER=$(gh api "repos/$SOURCE_REPO/actions/runs/$SOURCE_RUN_ID" --jq '.pull_requests[0].number // empty')
-          if [ -z "$PR_NUMBER" ]; then
-            echo "No pull request is associated with CI run $SOURCE_RUN_ID."
-            exit 0
-          fi
-
-          ARTIFACT_NAME=$(gh api "repos/$SOURCE_REPO/actions/runs/$SOURCE_RUN_ID/artifacts" \
-            --jq '.artifacts[] | select(.expired == false and (.name | startswith("e2e-evidence-"))) | .name' \
-            | python3 -c 'import sys; print(next(iter(sys.stdin), "").strip())')
-          if [ -z "$ARTIFACT_NAME" ]; then
-            echo "No E2E evidence artifact was produced for CI run $SOURCE_RUN_ID."
-            exit 0
-          fi
-
-          EVIDENCE_DIR="$RUNNER_TEMP/e2e-evidence"
-          mkdir -p "$EVIDENCE_DIR"
-          gh run download "$SOURCE_RUN_ID" --repo "$SOURCE_REPO" --name "$ARTIFACT_NAME" --dir "$EVIDENCE_DIR"
-
-          python3 scripts/ci/publish_e2e_evidence.py \
-            --evidence-dir "$EVIDENCE_DIR" \
-            --source-repo "$SOURCE_REPO" \
-            --pr-number "$PR_NUMBER"

.github/workflows/review-labels.yml

@@ -1,109 +0,0 @@
-name: Review labels
-
-# Require explicit maintainer review when CI-sensitive files or the MCP
-# catalog change. Previously this was split across two jobs in two
-# workflows: ``ci-review`` in lint.yml (gated on ``ci_review``) and
-# ``mcp-catalog-review`` in supply-chain-audit.yml (gated on
-# ``mcp_catalog``). Both checked for their own label.
-#
-# Now consolidated: a single ``ci-reviewed`` label covers both. The
-# comment sections tell the reviewer exactly what to verify per area,
-# so one label is enough — the human reads the comment, not the label
-# name.
-#
-# Outputs:
-#   ci_reviewed    — "true" / "false" / "" (empty when neither lane ran)
-#   review_status  — JSON array of status objects consumed by the review
-#                    comment assembler. See scripts/ci/emit_review_status.py.
-
-on:
-  workflow_call:
-    inputs:
-      ci_review:
-        description: Whether CI-sensitive files (eslint config, workflows, actions) changed.
-        type: boolean
-        default: false
-      ci_review_files:
-        description: JSON list of CI-sensitive files changed by the pull request.
-        type: string
-        default: '[]'
-      mcp_catalog:
-        description: Whether the MCP catalog / installer changed.
-        type: boolean
-        default: false
-      supply_chain:
-        description: Whether the critical supply-chain scan found a risk requiring review.
-        type: boolean
-        default: false
-    outputs:
-      ci_reviewed:
-        description: Whether the ci-reviewed label is present. Empty when neither input was true.
-        value: ${{ jobs.check.outputs.ci_reviewed }}
-      review_status:
-        description: JSON array of status objects for the review comment assembler.
-        value: ${{ jobs.check.outputs.review_status }}
-
-permissions:
-  contents: read
-  pull-requests: read # read PR labels
-
-jobs:
-  check:
-    name: Review label gate
-    if: inputs.ci_review || inputs.mcp_catalog || inputs.supply_chain
-    runs-on: ubuntu-latest
-    timeout-minutes: 2
-    outputs:
-      ci_reviewed: ${{ steps.label-check.outputs.ci_reviewed }}
-      review_status: ${{ steps.build-status.outputs.review_status }}
-    steps:
-      - name: Checkout code
-        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
-
-      - name: Check ci-reviewed label
-        id: label-check
-        env:
-          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-          REPO: ${{ github.repository }}
-        run: |
-          set -euo pipefail
-          PR="${{ github.event.pull_request.number }}"
-          LABELS=$(gh pr view "$PR" --repo "$REPO" --json labels --jq '.labels[].name' || true)
-
-          if echo "$LABELS" | grep -Fxq 'ci-reviewed'; then
-            echo "ci-reviewed label present."
-            echo "ci_reviewed=true" >> "$GITHUB_OUTPUT"
-          else
-            echo "ci-reviewed label missing."
-            echo "ci_reviewed=false" >> "$GITHUB_OUTPUT"
-          fi
-
-      - name: Build review_status JSON
-        id: build-status
-        env:
-          CI_REVIEW: ${{ inputs.ci_review }}
-          CI_REVIEW_FILES: ${{ inputs.ci_review_files }}
-          MCP_CATALOG: ${{ inputs.mcp_catalog }}
-          SUPPLY_CHAIN: ${{ inputs.supply_chain }}
-          LABEL_PRESENT: ${{ steps.label-check.outputs.ci_reviewed }}
-          REPO_URL: ${{ github.server_url }}/${{ github.repository }}
-          BASE_SHA: ${{ github.event.pull_request.base.sha }}
-          HEAD_SHA: ${{ github.event.pull_request.head.sha }}
-        run: |
-          set -euo pipefail
-          args=()
-          if [ "$CI_REVIEW" = "true" ]; then args+=(--ci-review); fi
-          args+=(--ci-review-files "$CI_REVIEW_FILES")
-          if [ "$MCP_CATALOG" = "true" ]; then args+=(--mcp-catalog); fi
-          if [ "$SUPPLY_CHAIN" = "true" ]; then args+=(--supply-chain); fi
-          if [ "$LABEL_PRESENT" = "true" ]; then args+=(--label-present); fi
-
-          python3 scripts/ci/emit_review_status.py "${args[@]}" \
-            --repo-url "$REPO_URL" --base-sha "$BASE_SHA" --head-sha "$HEAD_SHA" \
-            --output "$GITHUB_OUTPUT"
-
-      - name: Fail on missing label
-        if: steps.label-check.outputs.ci_reviewed != 'true'
-        run: |
-          echo "::error::CI-sensitive changes require the ci-reviewed label. Add the label and re-run this check."
-          exit 1

.github/workflows/skills-index-freshness.yml

@@ -20,8 +20,6 @@ jobs:
   check-freshness:
     if: github.repository == 'NousResearch/hermes-agent'
     runs-on: ubuntu-latest
-    timeout-minutes: 10
-    environment: trusted-automation
     steps:
       - name: Probe live index
         id: probe
@@ -30,7 +28,7 @@ jobs:
           URL="https://hermes-agent.nousresearch.com/docs/api/skills-index.json"
           echo "Probing $URL"
           # -L follows redirects; -f fails on HTTP errors; -s suppresses progress
-          if ! curl -fsSL --retry 3 --retry-delay 10 -o /tmp/skills-index.json "$URL"; then
+          if ! curl -fsSL -o /tmp/skills-index.json "$URL"; then
             echo "status=fetch-failed" >> "$GITHUB_OUTPUT"
             echo "detail=Could not download $URL" >> "$GITHUB_OUTPUT"
             exit 0
@@ -109,18 +107,10 @@ jobs:
             echo "Summary:      ${{ steps.probe.outputs.summary }}"
           fi
 
-      - name: Get GitHub App token
-        if: steps.probe.outputs.status != 'ok'
-        id: app-token
-        uses: ./.github/actions/get-app-token
-        with:
-          client-id: ${{ vars.APP_CLIENT_ID }}
-          private-key: ${{ secrets.APP_PRIVATE_KEY }}
-
       - name: Open issue on degraded / failed probe
         if: steps.probe.outputs.status != 'ok'
         env:
-          GH_TOKEN: ${{ steps.app-token.outputs.token }}
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
           STATUS: ${{ steps.probe.outputs.status }}
           DETAIL: ${{ steps.probe.outputs.detail }}
         run: |

.github/workflows/skills-index.yml

@@ -3,51 +3,40 @@ name: Build Skills Index
 on:
   schedule:
     # Run twice daily: 6 AM and 6 PM UTC
-    - cron: "0 6,18 * * *"
-  workflow_dispatch: # Manual trigger
+    - cron: '0 6,18 * * *'
+  workflow_dispatch:  # Manual trigger
   push:
     branches: [main]
     paths:
-      - "scripts/build_skills_index.py"
-      - ".github/workflows/skills-index.yml"
+      - 'scripts/build_skills_index.py'
+      - '.github/workflows/skills-index.yml'
 
 permissions:
   contents: read
-  actions: write # to trigger deploy-site.yml on schedule
+  actions: write   # to trigger deploy-site.yml on schedule
 
 jobs:
   build-index:
     # Only run on the upstream repository, not on forks
     if: github.repository == 'NousResearch/hermes-agent'
     runs-on: ubuntu-latest
-    timeout-minutes: 15
-    environment: trusted-automation
     steps:
-      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
+      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd  # v6.0.2
 
-      - name: Get GitHub App token
-        id: app-token
-        uses: ./.github/actions/get-app-token
+      - uses: actions/setup-python@a309ff8b426b58ec0e2a45f0f869d46889d02405  # v6.2.0
         with:
-          client-id: ${{ vars.APP_CLIENT_ID }}
-          private-key: ${{ secrets.APP_PRIVATE_KEY }}
-
-      - uses: actions/setup-python@a309ff8b426b58ec0e2a45f0f869d46889d02405 # v6.2.0
-        with:
-          python-version: "3.11"
+          python-version: '3.11'
 
       - name: Install dependencies
-        uses: ./.github/actions/retry
-        with:
-          command: pip install httpx==0.28.1 pyyaml==6.0.2
+        run: pip install httpx==0.28.1 pyyaml==6.0.2
 
       - name: Build skills index
         env:
-          GITHUB_TOKEN: ${{ steps.app-token.outputs.token }}
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
         run: python scripts/build_skills_index.py
 
       - name: Upload index artifact
-        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02  # v4
         with:
           name: skills-index
           path: website/static/api/skills-index.json
@@ -60,16 +49,8 @@ jobs:
     needs: build-index
     if: github.event_name == 'schedule' || github.event_name == 'workflow_dispatch'
     runs-on: ubuntu-latest
-    timeout-minutes: 15
-    environment: trusted-automation
     steps:
-      - name: Get GitHub App token
-        id: app-token
-        uses: ./.github/actions/get-app-token
-        with:
-          client-id: ${{ vars.APP_CLIENT_ID }}
-          private-key: ${{ secrets.APP_PRIVATE_KEY }}
       - name: Trigger Deploy Site workflow
         env:
-          GH_TOKEN: ${{ steps.app-token.outputs.token }}
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
         run: gh workflow run deploy-site.yml --repo ${{ github.repository }} -f skills_index_run_id=${{ github.run_id }}

.github/workflows/supply-chain-audit.yml

@@ -1,5 +1,16 @@
 name: Supply Chain Audit
 
+on:
+  # No paths filter — the jobs must always run so required checks
+  # report a status (path-gated workflows leave checks "pending" forever
+  # when no matching files change, which blocks merge).
+  pull_request:
+    types: [opened, synchronize, reopened]
+
+permissions:
+  pull-requests: write
+  contents: read
+
 # Narrow, high-signal scanner. Only fires on critical indicators of supply
 # chain attacks (e.g. the litellm-style payloads). Low-signal heuristics
 # (plain base64, plain exec/eval, dependency/Dockerfile/workflow edits,
@@ -8,57 +19,57 @@ name: Supply Chain Audit
 # the scanner. Keep this file's checks ruthlessly narrow: if you find
 # yourself adding WARNING-tier patterns here again, make a separate
 # advisory-only workflow instead.
-#
-# Path-gating is handled centrally by the ``ci.yml`` orchestrator's
-# ``detect`` job. The orchestrator passes ``scan`` / ``deps`` booleans as
-# inputs; this workflow's jobs gate on those inputs instead of re-computing
-# the diff. MCP catalog review was previously here but has moved to
-# ``review-labels.yml`` so it can be rerun independently.
-#
-# Outputs:
-#   review_status      — JSON array of status objects consumed by the review
-#                         comment assembler (scripts/ci/assemble_review_comment.py).
-#   critical_findings  — "true" when the narrow critical-pattern scan found
-#                         something. The review-label gate consumes this and
-#                         owns the action-required result, so adding
-#                         ``ci-reviewed`` can heal the run on rerun.
-
-on:
-  workflow_call:
-    inputs:
-      event_name:
-        description: The event name from the calling orchestrator.
-        type: string
-        required: true
-      scan:
-        description: Whether supply-chain-relevant files changed.
-        type: boolean
-        required: true
-      deps:
-        description: Whether pyproject.toml changed.
-        type: boolean
-        required: true
-    outputs:
-      review_status:
-        description: JSON array of review status objects for the review comment assembler.
-        value: ${{ jobs.aggregate.outputs.review_status }}
-      critical_findings:
-        description: Whether the critical-pattern scan found a risk requiring maintainer review.
-        value: ${{ jobs.aggregate.outputs.critical_findings }}
-
-permissions:
-  pull-requests: write
-  contents: read
 
 jobs:
+  # ── Path filter (shared by both scan and dep-bounds) ───────────────
+  changes:
+    runs-on: ubuntu-latest
+    outputs:
+      # True when any file the scanner cares about changed in this PR
+      scan: ${{ steps.filter.outputs.scan }}
+      # True when pyproject.toml changed in this PR
+      deps: ${{ steps.filter.outputs.deps }}
+      # True when the curated MCP catalog / bundled MCP manifests changed.
+      mcp_catalog: ${{ steps.filter.outputs.mcp_catalog }}
+    steps:
+      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
+        with:
+          fetch-depth: 0
+      - name: Check for relevant file changes
+        id: filter
+        run: |
+          BASE="${{ github.event.pull_request.base.sha }}"
+          HEAD="${{ github.event.pull_request.head.sha }}"
+          SCAN_FILES=$(git diff --name-only "$BASE"..."$HEAD" -- \
+            '*.py' '**/*.py' '*.pth' '**/*.pth' \
+            'setup.py' 'setup.cfg' \
+            'sitecustomize.py' 'usercustomize.py' '__init__.pth' \
+            'pyproject.toml' || true)
+          if [ -n "$SCAN_FILES" ]; then
+            echo "scan=true" >> "$GITHUB_OUTPUT"
+          else
+            echo "scan=false" >> "$GITHUB_OUTPUT"
+          fi
+          DEPS_FILES=$(git diff --name-only "$BASE"..."$HEAD" -- 'pyproject.toml' || true)
+          if [ -n "$DEPS_FILES" ]; then
+            echo "deps=true" >> "$GITHUB_OUTPUT"
+          else
+            echo "deps=false" >> "$GITHUB_OUTPUT"
+          fi
+          MCP_CATALOG_FILES=$(git diff --name-only "$BASE"..."$HEAD" -- \
+            'optional-mcps/**' \
+            'hermes_cli/mcp_catalog.py' || true)
+          if [ -n "$MCP_CATALOG_FILES" ]; then
+            echo "mcp_catalog=true" >> "$GITHUB_OUTPUT"
+          else
+            echo "mcp_catalog=false" >> "$GITHUB_OUTPUT"
+          fi
+
   scan:
     name: Scan PR for critical supply chain risks
-    if: inputs.scan
+    needs: changes
+    if: needs.changes.outputs.scan == 'true'
     runs-on: ubuntu-latest
-    timeout-minutes: 15
-    outputs:
-      review_status: ${{ steps.emit-status.outputs.review_status }}
-      critical_findings: ${{ steps.scan.outputs.found }}
     steps:
       - name: Checkout
         uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
@@ -68,8 +79,7 @@ jobs:
       - name: Scan diff for critical patterns
         id: scan
         env:
-          GH_TOKEN: ${{ github.token }}
-          CI_REVIEWED: ${{ contains(github.event.pull_request.labels.*.name, 'ci-reviewed') }}
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
         run: |
           set -euo pipefail
 
@@ -77,7 +87,7 @@ jobs:
           HEAD="${{ github.event.pull_request.head.sha }}"
 
           # Added lines only, excluding lockfiles.
-          # Three-point diff (base...head) diffs from the merge base to HEAD,
+          # Three-dot diff (base...head) diffs from the merge base to HEAD,
           # so only changes introduced by this PR are included — not changes
           # that landed on main after the PR branched off.
           DIFF=$(git diff "$BASE"..."$HEAD" -- . ':!uv.lock' ':!*.lock' ':!package-lock.json' ':!yarn.lock' || true)
@@ -87,7 +97,7 @@ jobs:
           # --- .pth files (auto-execute on Python startup) ---
           # The exact mechanism used in the litellm supply chain attack:
           # https://github.com/BerriAI/litellm/issues/24512
-          PTH_FILES=$(git diff --diff-filter=d --name-only "$BASE"..."$HEAD" | grep '\.pth$' || true)
+          PTH_FILES=$(git diff --name-only "$BASE"..."$HEAD" | grep '\.pth$' || true)
           if [ -n "$PTH_FILES" ]; then
             FINDINGS="${FINDINGS}
           ### 🚨 CRITICAL: .pth file added or modified
@@ -101,7 +111,7 @@ jobs:
           fi
 
           # --- base64 decode + exec/eval on the same line (the litellm attack pattern) ---
-          B64_EXEC_HITS=$(echo "$DIFF" | grep -n '^+' | grep -iE 'base64\.(b64decode|decodebytes|urlsafe_b64decode)' | grep -iE 'exec\(|eval\(' | head -10 || true)
+          B64_EXEC_HITS=$(echo "$DIFF" | grep -n '^\+' | grep -iE 'base64\.(b64decode|decodebytes|urlsafe_b64decode)' | grep -iE 'exec\(|eval\(' | head -10 || true)
           if [ -n "$B64_EXEC_HITS" ]; then
             FINDINGS="${FINDINGS}
           ### 🚨 CRITICAL: base64 decode + exec/eval combo
@@ -115,7 +125,7 @@ jobs:
           fi
 
           # --- subprocess with encoded/obfuscated command argument ---
-          PROC_HITS=$(echo "$DIFF" | grep -n '^+' | grep -E 'subprocess\.(Popen|call|run)\s*\(' | grep -iE 'base64|\\x[0-9a-f]{2}|chr\(' | head -10 || true)
+          PROC_HITS=$(echo "$DIFF" | grep -n '^\+' | grep -E 'subprocess\.(Popen|call|run)\s*\(' | grep -iE 'base64|\\x[0-9a-f]{2}|chr\(' | head -10 || true)
           if [ -n "$PROC_HITS" ]; then
             FINDINGS="${FINDINGS}
           ### 🚨 CRITICAL: subprocess with encoded/obfuscated command
@@ -135,11 +145,8 @@ jobs:
           # auto-loaded by the interpreter via site.py. Any nested file with the
           # same name (e.g. hermes_cli/setup.py — the CLI setup wizard) is unrelated
           # and produced false positives that trained reviewers to ignore the scanner.
-          SETUP_HITS=$(git diff --diff-filter=d --name-only "$BASE"..."$HEAD" | grep -E '^(setup\.py|setup\.cfg|sitecustomize\.py|usercustomize\.py|__init__\.pth)$' || true)
-          # A maintainer-applied ci-reviewed label records the manual review
-          # required for intentional changes to an install hook. The scanner
-          # still blocks every unreviewed addition or modification.
-          if [ -n "$SETUP_HITS" ] && [ "$CI_REVIEWED" != "true" ]; then
+          SETUP_HITS=$(git diff --name-only "$BASE"..."$HEAD" | grep -E '^(setup\.py|setup\.cfg|sitecustomize\.py|usercustomize\.py|__init__\.pth)$' || true)
+          if [ -n "$SETUP_HITS" ]; then
             FINDINGS="${FINDINGS}
           ### 🚨 CRITICAL: Install-hook file added or modified
           These files can execute code during package installation or interpreter startup.
@@ -158,32 +165,46 @@ jobs:
             echo "found=false" >> "$GITHUB_OUTPUT"
           fi
 
-      - name: Emit review_status
-        id: emit-status
-        if: always()
+      - name: Post critical finding comment
+        if: steps.scan.outputs.found == 'true'
         env:
-          FOUND: ${{ steps.scan.outputs.found }}
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
         run: |
-          python3 - <<'PYEOF'
-          import json, os
+          BODY="## 🚨 CRITICAL Supply Chain Risk Detected
 
-          # The review-label gate renders and blocks critical findings. Keep
-          # this scan a fact-finder so adding ci-reviewed can rerun the gate
-          # without requiring the scanner itself to fail again.
-          status = []
+          This PR contains a pattern that has been used in real supply chain attacks. A maintainer must review the flagged code carefully before merging.
 
-          with open(os.environ["GITHUB_OUTPUT"], "a", encoding="utf-8") as f:
-              f.write(f"review_status={json.dumps(status)}\n")
-          PYEOF
+          $(cat /tmp/findings.md)
 
+          ---
+          *Scanner only fires on high-signal indicators: .pth files, base64+exec/eval combos, subprocess with encoded commands, or install-hook files. Low-signal warnings were removed intentionally — if you're seeing this comment, the finding is worth inspecting.*"
+
+          gh pr comment "${{ github.event.pull_request.number }}" --body "$BODY" || echo "::warning::Could not post PR comment (expected for fork PRs — GITHUB_TOKEN is read-only)"
+
+      - name: Fail on critical findings
+        if: steps.scan.outputs.found == 'true'
+        run: |
+          echo "::error::CRITICAL supply chain risk patterns detected in this PR. See the PR comment for details."
+          exit 1
+
+  # Gate: reports success when scan was skipped (no relevant files changed).
+  # This ensures the required check always gets a status.
+  scan-gate:
+    name: Scan PR for critical supply chain risks
+    needs: changes
+    # always() so the gate still reports SUCCESS even if `changes` fails/is
+    # skipped — without it, a failed dependency would leave the required
+    # check unreported (i.e. "pending"), the exact failure mode this fixes.
+    if: always() && needs.changes.outputs.scan != 'true'
+    runs-on: ubuntu-latest
+    steps:
+      - run: echo "No supply-chain-relevant files changed, skipping scan."
 
   dep-bounds:
     name: Check PyPI dependency upper bounds
-    if: inputs.deps
+    needs: changes
+    if: needs.changes.outputs.deps == 'true'
     runs-on: ubuntu-latest
-    timeout-minutes: 15
-    outputs:
-      review_status: ${{ steps.emit-status.outputs.review_status }}
     steps:
       - name: Checkout
         uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
@@ -206,7 +227,7 @@ jobs:
             exit 0
           fi
 
-          # Match PyPI dep specs that have >= and no < ceiling.
+          # Match PyPI dep specs that have >= but no < ceiling.
           # Pattern: "package>=version" without a following ",<" bound.
           # Excludes git+ URLs (which use commit SHAs) and comments.
           UNBOUNDED=$(echo "$ADDED" | grep -oE '"[a-zA-Z0-9_-]+(\[[^\]]*\])?>=[ 0-9.]+"' | grep -v ',<' || true)
@@ -218,36 +239,26 @@ jobs:
             echo "found=false" >> "$GITHUB_OUTPUT"
           fi
 
-      - name: Emit review_status
-        id: emit-status
-        if: always()
+      - name: Post unbounded dep warning
+        if: steps.bounds.outputs.found == 'true'
         env:
-          FOUND: ${{ steps.bounds.outputs.found }}
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
         run: |
-          python3 - <<'PYEOF'
-          import json, os
+          BODY="## ⚠️ Unbounded PyPI Dependency Detected
 
-          found = os.environ.get("FOUND", "") == "true"
+          This PR adds PyPI dependencies without a \`<next_major\` upper bound. Per our [supply chain policy](../blob/main/CONTRIBUTING.md#dependency-pinning-policy-supply-chain-hardening), all PyPI deps must be pinned as \`>=floor,<next_major\`.
 
-          if found:
-              with open("/tmp/unbounded.txt", encoding="utf-8") as f:
-                  detail = f.read()
-              status = [{
-                  "source": "supply chain",
-                  "results": [{
-                      "kind": "action_required",
-                      "title": "Unbounded PyPI dependencies",
-                      "summary": "This PR adds PyPI dependencies without upper bounds.",
-                      "detail": detail,
-                      "how_to_fix": 'Add a `<next_major` upper bound, e.g. `"package>=1.2.0,<2"`. See CONTRIBUTING.md dependency pinning policy.'
-                  }]
-              }]
-          else:
-              status = []
+          **Unbounded specs found:**
+          \`\`\`
+          $(cat /tmp/unbounded.txt)
+          \`\`\`
 
-          with open(os.environ["GITHUB_OUTPUT"], "a", encoding="utf-8") as f:
-              f.write(f"review_status={json.dumps(status)}\n")
-          PYEOF
+          **Fix:** Add an upper bound, e.g. \`\"package>=1.2.0,<2\"\`
+
+          ---
+          *See PR #2810 and CONTRIBUTING.md for the full policy rationale.*"
+
+          gh pr comment "${{ github.event.pull_request.number }}" --body "$BODY" || echo "::warning::Could not post PR comment (expected for fork PRs)"
 
       - name: Fail on unbounded deps
         if: steps.bounds.outputs.found == 'true'
@@ -255,39 +266,62 @@ jobs:
           echo "::error::PyPI dependencies without upper bounds detected. Add <next_major ceiling per CONTRIBUTING.md policy."
           exit 1
 
-  aggregate:
-    name: Aggregate review statuses
-    needs: [scan, dep-bounds]
-    if: always()
+  # Gate: reports success when dep-bounds was skipped (no pyproject.toml changed).
+  # This ensures the required check always gets a status.
+  dep-bounds-gate:
+    name: Check PyPI dependency upper bounds
+    needs: changes
+    # always() so the gate still reports SUCCESS even if `changes` fails/is
+    # skipped — without it, a failed dependency would leave the required
+    # check unreported (i.e. "pending"), the exact failure mode this fixes.
+    if: always() && needs.changes.outputs.deps != 'true'
     runs-on: ubuntu-latest
-    timeout-minutes: 15
-    outputs:
-      review_status: ${{ steps.merge.outputs.review_status }}
-      critical_findings: ${{ steps.merge.outputs.critical_findings }}
     steps:
-      - name: Merge review statuses
-        id: merge
+      - run: echo "No pyproject.toml changes, skipping dependency bounds check."
+
+  mcp-catalog-review:
+    name: MCP catalog security review
+    needs: changes
+    if: needs.changes.outputs.mcp_catalog == 'true'
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
+        with:
+          fetch-depth: 0
+
+      - name: Require explicit MCP catalog review label
         env:
-          SCAN_STATUS: ${{ needs.scan.outputs.review_status }}
-          DEP_STATUS: ${{ needs.dep-bounds.outputs.review_status }}
-          CRITICAL_FINDINGS: ${{ needs.scan.outputs.critical_findings }}
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
         run: |
-          python3 - <<'PYEOF'
-          import json, os
+          set -euo pipefail
+          PR="${{ github.event.pull_request.number }}"
+          LABELS=$(gh pr view "$PR" --json labels --jq '.labels[].name' || true)
+          if echo "$LABELS" | grep -Fxq 'mcp-catalog-reviewed'; then
+            echo "MCP catalog review label present."
+            exit 0
+          fi
 
-          merged = []
-          for key in ("SCAN_STATUS", "DEP_STATUS"):
-              raw = os.environ.get(key, "")
-              if not raw:
-                  continue
-              try:
-                  data = json.loads(raw)
-              except (json.JSONDecodeError, TypeError):
-                  continue
-              if isinstance(data, list):
-                  merged.extend(data)
+          BODY="## ⚠️ MCP catalog security review required
 
-          with open(os.environ["GITHUB_OUTPUT"], "a", encoding="utf-8") as f:
-              f.write(f"review_status={json.dumps(merged)}\n")
-              f.write("critical_findings=" + os.environ.get("CRITICAL_FINDINGS", "false") + "\n")
-          PYEOF
+          This PR changes the bundled MCP catalog or MCP catalog installer code. MCP entries can define local commands that users later install into \`mcp_servers\`, so this needs explicit maintainer review before merge.
+
+          A maintainer should verify:
+          - any new/changed \`optional-mcps/**/manifest.yaml\` command and args are expected,
+          - stdio transports do not use shell+egress/exfiltration payloads,
+          - git install refs are pinned and bootstrap commands are minimal,
+          - requested env vars/secrets match the upstream MCP's documented needs.
+
+          After review, add the \`mcp-catalog-reviewed\` label and re-run this check."
+
+          gh pr comment "$PR" --body "$BODY" || echo "::warning::Could not post PR comment (expected for fork PRs)"
+          echo "::error::MCP catalog changes require the mcp-catalog-reviewed label."
+          exit 1
+
+  mcp-catalog-review-gate:
+    name: MCP catalog security review
+    needs: changes
+    if: always() && needs.changes.outputs.mcp_catalog != 'true'
+    runs-on: ubuntu-latest
+    steps:
+      - run: echo "No MCP catalog changes, skipping MCP catalog security review."

.github/workflows/tests.yml

@@ -1,28 +1,33 @@
 name: Tests
 
 on:
-  workflow_call:
-    inputs:
-      slice_count:
-        description: Number of parallel test slices
-        type: number
-        default: 8
+  push:
+    branches: [main]
+    paths-ignore:
+      - "**/*.md"
+      - "docs/**"
+  # No paths filter — the job must always run so the required check
+  # reports a status (path-gated workflows leave checks "pending" forever
+  # when no matching files change, which blocks merge).
+  pull_request:
+    branches: [main]
 
 permissions:
   contents: read
 
-# Cancel in-progress runs for the same ref
+# Cancel in-progress runs for the same PR/branch
 concurrency:
   group: tests-${{ github.ref }}
   cancel-in-progress: true
 
 jobs:
-  generate:
-    name: "Generate slices"
+  test:
     runs-on: ubuntu-latest
-    timeout-minutes: 10
-    outputs:
-      matrix: ${{ steps.matrix.outputs.matrix }}
+    timeout-minutes: 30
+    strategy:
+      fail-fast: false
+      matrix:
+        slice: [1, 2, 3, 4, 5, 6]
     steps:
       - name: Checkout code
         uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
@@ -31,31 +36,12 @@ jobs:
         uses: actions/cache/restore@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
         with:
           path: test_durations.json
+          # main always writes a new suffix, but jobs pick the latest one with the same prefix
+          # quote from https://docs.github.com/en/actions/reference/workflows-and-actions/dependency-caching#cache-hits-and-misses
+          # If you provide restore-keys, the cache action sequentially searches for any caches that match the list of restore-keys.
+          # If there are no exact matches, the action searches for partial matches of the restore keys.
+          # When the action finds a partial match, the most recent cache is restored to the path directory.
           key: test-durations
-          # Saves use test-durations-${run_id}, so the exact key above never
-          # matches — without this prefix fallback the cache ALWAYS missed,
-          # LPT slicing ran on no data, and unbalanced slices pushed heavy
-          # files toward the per-file timeout under load.
-          restore-keys: |
-            test-durations-
-
-      - name: Generate test slices
-        id: matrix
-        run: |
-          MATRIX=$(python3 scripts/run_tests_parallel.py --generate-slices ${{ inputs.slice_count }})
-          echo "matrix=$MATRIX" >> "$GITHUB_OUTPUT"
-
-  test:
-    name: Run tests slice ${{ matrix.slice.index }}/${{ inputs.slice_count }}
-    needs: generate
-    runs-on: ubuntu-latest
-    timeout-minutes: 30
-    strategy:
-      fail-fast: false
-      matrix: ${{ fromJSON(needs.generate.outputs.matrix) }}
-    steps:
-      - name: Checkout code
-        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
 
       - name: Install ripgrep (prebuilt binary)
         run: |
@@ -63,7 +49,7 @@ jobs:
           RG_VERSION=15.1.0
           RG_SHA256=1c9297be4a084eea7ecaedf93eb03d058d6faae29bbc57ecdaf5063921491599
           RG_TARBALL=ripgrep-${RG_VERSION}-x86_64-unknown-linux-musl.tar.gz
-          curl -sSfL --retry 3 --retry-delay 5 -o "$RG_TARBALL" \
+          curl -sSfL -o "$RG_TARBALL" \
             "https://github.com/BurntSushi/ripgrep/releases/download/${RG_VERSION}/${RG_TARBALL}"
           echo "${RG_SHA256}  ${RG_TARBALL}" | sha256sum -c -
           tar -xzf "$RG_TARBALL"
@@ -72,7 +58,7 @@ jobs:
           rg --version
 
       - name: Install uv
-        uses: astral-sh/setup-uv@fac544c07dec837d0ccb6301d7b5580bf5edae39 # 8.2.0
+        uses: astral-sh/setup-uv@d4b2f3b6ecc6e67c4457f6d3e41ec42d3d0fcb86 # v5
         with:
           # Persist uv's download/wheel cache (~/.cache/uv) across runs.
           # Keyed on the dependency manifests, so the cache is reused until
@@ -92,28 +78,40 @@ jobs:
         # fails if the lock is out of sync with pyproject.toml), giving a
         # reproducible env. It also creates .venv itself, so no separate
         # `uv venv` step is needed.
-        uses: ./.github/actions/retry
-        with:
-          command: uv sync --locked --python 3.11 --extra all --extra dev
+        run: uv sync --locked --python 3.11 --extra all --extra dev
 
       - name: Minimize uv cache
         # Optimized for CI: prunes pre-built wheels that are cheap to
         # re-download, keeping the persisted cache small and fast to restore.
         run: uv cache prune --ci
 
-      - name: Run tests (slice ${{ matrix.slice.index }}/${{ inputs.slice_count }})
-        # Per-file isolation via scripts/run_tests.sh: each test file runs
-        # in its own freshly-spawned `python -m pytest <file>` subprocess
+      - name: Run tests (slice ${{ matrix.slice }}/6)
+        # Per-file isolation via scripts/run_tests_parallel.py: discovers
+        # every test_*.py file under tests/ (excluding integration/ + e2e/),
+        # then runs `python -m pytest <file>` in a freshly-spawned subprocess
         # with bounded parallelism. No xdist, no shared workers, no
         # module-level state leakage between files.
         #
-        # File list is pre-computed by the generate job (--generate-slices)
-        # which runs LPT distribution once and passes the file list to each
-        # matrix job via --files. Previously each job re-discovered files and
-        # re-ran LPT independently — redundant N times.
+        # Why per-file (not per-test): per-test spawn cost (~250ms × 17k
+        # tests = 70min CPU minimum) blew the wall-clock budget. Per-file
+        # spawn (~250ms × ~850 files = ~3.5min) fits while still giving
+        # every file a fresh interpreter — the only isolation boundary
+        # that matters in practice (cross-file leakage was the original
+        # flake source; intra-file is the test author's responsibility).
+        #
+        # Why drop xdist entirely: xdist's persistent workers accumulate
+        # state across files, which is exactly the leakage we wanted to
+        # fix. ThreadPoolExecutor + subprocess.run is ~60 lines and does
+        # the job with cleaner semantics.
+        #
+        # Matrix slicing (--slice I/N): files are distributed across 6
+        # jobs by cached duration (LPT algorithm) so each job gets
+        # roughly equal wall time. Without a cache, files default to 2s
+        # estimate and get split roughly evenly by count — still correct,
+        # just not perfectly balanced.
         run: |
           source .venv/bin/activate
-          scripts/run_tests.sh --files '${{ matrix.slice.files }}'
+          python scripts/run_tests_parallel.py --slice ${{ matrix.slice }}/6
         env:
           # Ensure tests don't accidentally call real APIs
           OPENROUTER_API_KEY: ""
@@ -121,12 +119,9 @@ jobs:
           NOUS_API_KEY: ""
 
       - name: Upload per-slice durations
-        # Advisory artifact (feeds slice balancing) — a transient artifact-
-        # service blip must not fail an otherwise-green test slice.
-        continue-on-error: true
         uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
         with:
-          name: test-durations-slice-${{ matrix.slice.index }}
+          name: test-durations-slice-${{ matrix.slice }}
           path: test_durations.json
           retention-days: 1
 
@@ -136,7 +131,6 @@ jobs:
     needs: test
     if: needs.test.result == 'success' && github.ref == 'refs/heads/main'
     runs-on: ubuntu-latest
-    timeout-minutes: 10
     steps:
       - name: Download all slice durations
         uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1
@@ -177,7 +171,7 @@ jobs:
           RG_VERSION=15.1.0
           RG_SHA256=1c9297be4a084eea7ecaedf93eb03d058d6faae29bbc57ecdaf5063921491599
           RG_TARBALL=ripgrep-${RG_VERSION}-x86_64-unknown-linux-musl.tar.gz
-          curl -sSfL --retry 3 --retry-delay 5 -o "$RG_TARBALL" \
+          curl -sSfL -o "$RG_TARBALL" \
             "https://github.com/BurntSushi/ripgrep/releases/download/${RG_VERSION}/${RG_TARBALL}"
           echo "${RG_SHA256}  ${RG_TARBALL}" | sha256sum -c -
           tar -xzf "$RG_TARBALL"
@@ -186,7 +180,7 @@ jobs:
           rg --version
 
       - name: Install uv
-        uses: astral-sh/setup-uv@fac544c07dec837d0ccb6301d7b5580bf5edae39 # 8.2.0
+        uses: astral-sh/setup-uv@d4b2f3b6ecc6e67c4457f6d3e41ec42d3d0fcb86 # v5
         with:
           # Persist uv's download/wheel cache (~/.cache/uv) across runs.
           # Keyed on the dependency manifests, so the cache is reused until
@@ -206,15 +200,18 @@ jobs:
         # fails if the lock is out of sync with pyproject.toml), giving a
         # reproducible env. It also creates .venv itself, so no separate
         # `uv venv` step is needed.
-        uses: ./.github/actions/retry
-        with:
-          command: uv sync --locked --python 3.11 --extra all --extra dev
+        run: uv sync --locked --python 3.11 --extra all --extra dev
 
       - name: Minimize uv cache
         # Optimized for CI: prunes pre-built wheels that are cheap to
         # re-download, keeping the persisted cache small and fast to restore.
         run: uv cache prune --ci
 
+      - name: Packaged-wheel i18n smoke test
+        run: |
+          source .venv/bin/activate
+          python -m pytest -m integration tests/test_wheel_locales_e2e.py -v
+
       - name: Run e2e tests
         run: |
           source .venv/bin/activate

.github/workflows/typecheck.yml

@@ -0,0 +1,45 @@
+# .github/workflows/typecheck.yml
+name: Typecheck
+
+on:
+  push:
+    branches: [main]
+  # No paths filter — the job must always run so the required check
+  # reports a status (path-gated workflows leave checks "pending" forever
+  # when no matching files change, which blocks merge).
+  pull_request:
+    branches: [main]
+
+jobs:
+  typecheck:
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        package:
+          [ui-tui, web, apps/bootstrap-installer, apps/desktop, apps/shared]
+      fail-fast: false # report all failures, not just the first one
+    steps:
+      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4
+        with:
+          node-version: 22
+          cache: npm
+      - run: npm ci
+      - run: npm run --prefix ${{ matrix.package }} typecheck
+
+  # Production build of the desktop renderer. `typecheck` runs `tsc` only,
+  # which does NOT exercise Vite/Rolldown module resolution — so an
+  # unresolvable package export (e.g. a transitive @assistant-ui/tap that no
+  # longer exports "./react-shim") slips past typecheck and only explodes when
+  # users build apps/desktop from source on install/update. Run the real
+  # `vite build` here so that class of break fails in CI instead.
+  desktop-build:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4
+        with:
+          node-version: 22
+          cache: npm
+      - run: npm ci
+      - run: npm run --prefix apps/desktop build

.github/workflows/upload_to_pypi.yml

@@ -0,0 +1,164 @@
+name: Publish to PyPI
+
+# Triggered by CalVer tag pushes from scripts/release.py (e.g. v2026.5.15)
+# Can also be triggered manually from the Actions tab as an escape hatch.
+on:
+  push:
+    tags:
+      - 'v20*'  # CalVer tags: v2026.5.15, v2026.5.15.2, etc.
+  workflow_dispatch:
+    inputs:
+      confirm_tag:
+        description: 'Tag to publish (e.g. v2026.5.15). Must already exist.'
+        required: true
+        type: string
+
+# Restrict default token to read-only; each job escalates as needed.
+permissions:
+  contents: read
+
+# Prevent overlapping publishes (e.g. two same-day tags pushed quickly).
+concurrency:
+  group: pypi-publish
+  cancel-in-progress: false
+
+jobs:
+  build:
+    name: Build distribution 📦
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd  # v6.0.2
+        with:
+          persist-credentials: false
+          # On workflow_dispatch, check out the confirmed tag.
+          ref: ${{ inputs.confirm_tag || github.ref }}
+          fetch-tags: true
+
+      - name: Validate tag exists
+        if: github.event_name == 'workflow_dispatch'
+        run: |
+          if ! git tag -l "${{ inputs.confirm_tag }}" | grep -q .; then
+            echo "::error::Tag '${{ inputs.confirm_tag }}' does not exist in the repo"
+            exit 1
+          fi
+
+      - name: Set up Python
+        uses: actions/setup-python@a309ff8b426b58ec0e2a45f0f869d46889d02405  # v6.2.0
+        with:
+          python-version: '3.13'
+
+      - name: Install uv
+        uses: astral-sh/setup-uv@d0cc045d04ccac9d8b7881df0226f9e82c39688e  # v6
+
+      - name: Set up Node.js
+        uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020  # v4
+        with:
+          node-version: '22'
+
+      - name: Build web dashboard
+        run: cd web && npm ci && npm run build
+
+      - name: Build TUI bundle
+        run: cd ui-tui && npm ci && npm run build
+
+      - name: Bundle TUI into hermes_cli
+        run: |
+          mkdir -p hermes_cli/tui_dist
+          cp ui-tui/dist/entry.js hermes_cli/tui_dist/entry.js
+
+      - name: Verify frontend assets exist
+        run: |
+          test -f hermes_cli/web_dist/index.html || { echo "ERROR: web_dist not built"; exit 1; }
+          test -f hermes_cli/tui_dist/entry.js || { echo "ERROR: tui_dist not built"; exit 1; }
+
+      - name: Bundle install scripts into wheel
+        run: |
+          mkdir -p hermes_cli/scripts
+          cp scripts/install.sh hermes_cli/scripts/install.sh
+          cp scripts/install.ps1 hermes_cli/scripts/install.ps1
+
+      - name: Build wheel and sdist
+        run: uv build --sdist --wheel
+
+      - name: Upload distribution artifacts
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02  # v4
+        with:
+          name: python-package-distributions
+          path: dist/
+
+  publish:
+    name: Publish to PyPI
+    needs: build
+    runs-on: ubuntu-latest
+    environment:
+      name: pypi
+      url: https://pypi.org/p/hermes-agent
+    permissions:
+      id-token: write  # OIDC trusted publishing
+
+    steps:
+      - name: Download distribution artifacts
+        uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093  # v4
+        with:
+          name: python-package-distributions
+          path: dist/
+
+      - name: Publish to PyPI
+        uses: pypa/gh-action-pypi-publish@cef221092ed1bacb1cc03d23a2d87d1d172e277b  # v1.14.0
+        with:
+          skip-existing: true
+
+  sign:
+    name: Sign and attach to GitHub Release
+    # Only runs on tag pushes — release.py creates the GitHub Release,
+    # and workflow_dispatch won't have a matching release to attach to.
+    if: startsWith(github.ref, 'refs/tags/')
+    needs: publish
+    runs-on: ubuntu-latest
+    permissions:
+      contents: write   # attach assets to the existing release
+      id-token: write   # sigstore signing
+
+    steps:
+      - name: Download distribution artifacts
+        uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093  # v4
+        with:
+          name: python-package-distributions
+          path: dist/
+
+      - name: Wait for GitHub Release to exist
+        env:
+          GITHUB_TOKEN: ${{ github.token }}
+        # release.py creates the GitHub Release after pushing the tag,
+        # but this workflow starts from the tag push — wait for it.
+        run: |
+          for i in $(seq 1 30); do
+            if gh release view "$GITHUB_REF_NAME" --repo "$GITHUB_REPOSITORY" >/dev/null 2>&1; then
+              echo "Release $GITHUB_REF_NAME found"
+              exit 0
+            fi
+            echo "Waiting for release... ($i/30)"
+            sleep 10
+          done
+          echo "::warning::Release $GITHUB_REF_NAME not found after 5 minutes — skipping signature upload"
+          echo "skip_sign=true" >> "$GITHUB_ENV"
+
+      - name: Sign with Sigstore
+        if: env.skip_sign != 'true'
+        uses: sigstore/gh-action-sigstore-python@04cffa1d795717b140764e8b640de88853c92acc  # v3.3.0
+        with:
+          inputs: >-
+            ./dist/*.tar.gz
+            ./dist/*.whl
+
+      - name: Attach signed artifacts to GitHub Release
+        if: env.skip_sign != 'true'
+        env:
+          GITHUB_TOKEN: ${{ github.token }}
+        # release.py already created the GitHub Release — just upload
+        # the Sigstore signatures alongside the existing assets.
+        run: >-
+          gh release upload
+          "$GITHUB_REF_NAME" dist/*.sigstore.json
+          --repo "$GITHUB_REPOSITORY"
+          --clobber

.github/workflows/uv-lockfile-check.yml

@@ -4,7 +4,7 @@ name: uv.lock check
 # that modify pyproject.toml without regenerating uv.lock (or vice versa)
 # must not merge, because the Docker build's `uv sync --frozen` step will
 # fail on a stale lockfile and we'd rather catch it here than in the
-# docker workflow on main.
+# docker-publish workflow on main.
 #
 # ─────────────────────────────────────────────────────────────────────────
 # IMPORTANT: this check runs against the MERGED state, not just your branch
@@ -44,32 +44,37 @@ name: uv.lock check
 # the same way.  Better to catch it here than after merge.
 
 on:
-  workflow_call:
-    outputs:
-      review_status:
-        description: "JSON review status for the review-status aggregator"
-        value: ${{ jobs.check.outputs.review_status }}
+  push:
+    branches: [main]
+    paths:
+      - "pyproject.toml"
+      - "uv.lock"
+      - ".github/workflows/uv-lockfile-check.yml"
+
+  # No paths filter — the job must always run so the required check
+  # reports a status (path-gated workflows leave checks "pending" forever
+  # when no matching files change, which blocks merge).
+  pull_request:
+    branches: [main]
 
 permissions:
   contents: read
 
 concurrency:
   group: uv-lockfile-check-${{ github.event.pull_request.number || github.ref }}
-  cancel-in-progress: true
+  cancel-in-progress: ${{ github.event_name == 'pull_request' }}
 
 jobs:
   check:
     name: uv lock --check
     runs-on: ubuntu-latest
     timeout-minutes: 5
-    outputs:
-      review_status: ${{ steps.verify.outputs.review_status }}
     steps:
       - name: Checkout code
         uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
 
       - name: Install uv
-        uses: astral-sh/setup-uv@fac544c07dec837d0ccb6301d7b5580bf5edae39 # 8.2.0
+        uses: astral-sh/setup-uv@d4b2f3b6ecc6e67c4457f6d3e41ec42d3d0fcb86 # v5
 
       # `uv lock --check` re-resolves the project from pyproject.toml and
       # compares the result to uv.lock, exiting non-zero if they disagree.
@@ -79,22 +84,8 @@ jobs:
       # of this file) — failures often mean "your branch is behind main,
       # rebase and regenerate uv.lock."
       - name: Verify uv.lock is up-to-date
-        id: verify
         run: |
-          # uv lock --check re-resolves against PyPI (network). Retry so a
-          # registry blip doesn't read as "lockfile stale". A genuinely stale
-          # lockfile fails all attempts (deterministic), costing only seconds.
-          ok=false
-          for i in 1 2 3; do
-            if uv lock --check; then
-              ok=true
-              break
-            fi
-            [ "$i" = 3 ] && break
-            echo "::warning::uv lock --check failed (attempt $i); retrying in 10s"
-            sleep 10
-          done
-          if [ "$ok" != true ]; then
+          if ! uv lock --check; then
             cat <<'EOF' >> "$GITHUB_STEP_SUMMARY"
           ## ❌ uv.lock is out of sync with pyproject.toml
 
@@ -120,13 +111,9 @@ jobs:
 
           This check is blocking because the Docker image build uses
           `uv sync --frozen --extra all`, which rejects stale lockfiles
-          — catching it here avoids a ~15 min failed docker run
+          — catching it here avoids a ~15 min failed docker-publish run
           on `main` post-merge.
           EOF
             echo "::error title=uv.lock out of sync::Run \`uv lock\` locally and commit the result. If on a PR, sync with main first."
-            review_status='[{"source":"uv.lock check","results":[{"kind":"action_required","title":"uv.lock out of sync","summary":"uv.lock is out of sync with pyproject.toml.","how_to_fix":"Run `uv lock` locally and commit the result. If on a PR, sync with main first:\n```\ngit fetch origin main\ngit rebase origin/main\nuv lock\ngit add uv.lock\ngit commit -m \"chore: refresh uv.lock\"\n```\n"}]}]'
-            echo "review_status=${review_status}" >> "$GITHUB_OUTPUT"
             exit 1
           fi
-          review_status='[]'
-          echo "review_status=${review_status}" >> "$GITHUB_OUTPUT"

.gitignore

@@ -1,18 +1,13 @@
 .DS_Store
 /venv/
 /venv.old/
-/venv.stale.runtime-*/
-/.hermes-runtime/
 /_pycache/
 *.pyc*
 __pycache__/
-act/
-.act-sandbox-agent.*
 .venv/
 .venv
 .vscode/
 .env
-.op.env
 .env.local
 .env.development.local
 .env.test.local
@@ -46,10 +41,7 @@ run_datagen_sonnet.sh
 source-data/*
 run_datagen_megascience_glm4-6.sh
 data/*
-# No trailing slash: also matches node_modules SYMLINKS (worktrees often
-# symlink node_modules to the main checkout; the dir-only pattern let one
-# slip into a commit and break `npm ci` on CI with ENOTDIR).
-node_modules
+node_modules/
 browser-use/
 agent-browser/
 # Private keys
@@ -61,10 +53,6 @@ __pycache__/
 hermes_agent.egg-info/
 wandb/
 testlogs
-playwright-report/
-test-results/
-# Playwright visual regression baselines — cached from main in CI, not committed
-*-snapshots/
 
 # CLI config (may contain sensitive SSH paths)
 cli-config.yaml
@@ -77,23 +65,10 @@ environments/benchmarks/evals/
 
 # Web UI build output
 hermes_cli/web_dist/
-# Cross-process web UI build lock (flock target, always empty)
-.web_ui_build.lock
 apps/desktop/build/
 apps/desktop/dist/
-
-# tsc-emitted artifacts (a stray `tsc -b` compiles into src/, and vite then
-# resolves the stale .js OVER the .tsx — never track these)
-apps/desktop/src/**/*.js
-apps/desktop/src/**/*.js.map
-apps/desktop/src/**/*.d.ts
-!apps/desktop/src/global.d.ts
-!apps/desktop/src/vite-env.d.ts
-apps/shared/src/**/*.js
-apps/shared/src/**/*.js.map
-apps/shared/src/**/*.d.ts
 apps/desktop/release/
-*.tsbuildinfo
+apps/desktop/*.tsbuildinfo
 
 # Web UI assets — synced from @nous-research/ui at build time via
 # `npm run sync-assets` (see web/package.json).
@@ -143,20 +118,12 @@ docs/superpowers/*
 # treat it as a local edit and autostash it on every run (#38529).
 .hermes-bootstrap-complete
 
-# Persistent dev sandbox dir (scripts/dev-sandbox.sh --persistent)
-.hermes-sandbox/
-
 # Interrupted-update breadcrumb + recovery lock written next to the shared venv
 # by `hermes update` / launch-time self-heal. Runtime state, never a code change
 # — ignore so `git status` stays clean and update's autostash skips them.
 .update-incomplete
 .update-incomplete.lock
 
-# Installer-written method stamp in the managed checkout root (scripts/install.sh).
-# Runtime metadata only — never a code change. Ignore so `git status` stays clean
-# and `hermes update`'s untracked autostash does not treat it as a local edit (#66189 / #54855).
-/.install_method
-
 # Tool Search live-test harness output — non-deterministic model transcripts,
 # regenerated by scripts/tool_search_livetest.py. Never an artifact of the repo.
 scripts/out/
@@ -170,10 +137,3 @@ RELEASE_v*.md
 # Desktop demo-run scratch output (hermes writes demo/*.txt during recorded
 # walkthroughs). Throwaway artifacts, never part of the app.
 apps/desktop/demo/
-
-# PR infographics are rendered locally and embedded in PR descriptions via the
-# image-provider (fal.media) URL — they are NEVER committed to the repo. The
-# PR body is the archive. See the hermes-agent-dev skill's
-# pr-infographic-workflow reference (storage rule + lapse #8 / #COMMIT-1).
-infographic/
-native/fts5_cjk/*.so

.prettierignore

@@ -1,4 +0,0 @@
-# Lockfiles must never be reformatted — main has a repo rule requiring
-# team approval when lockfiles change, so an autofix PR touching one
-# would hang waiting for review.
-package-lock.json

AGENTS.md

@@ -123,17 +123,6 @@ conservative at the waist.
   without E2E proof, and plugins that touch core files.** Plugins live in their
   own directory and work within the ABCs/hooks we provide; if a plugin needs
   more, widen the generic plugin surface, don't special-case it in core.
-- **Third-party products / other people's projects integrated into the core
-  tree.** Observability backends, vendor SaaS integrations, analytics dashboards,
-  and similar "someone else's product" plugins do NOT land under `plugins/` in
-  this repo. They place an ongoing maintenance burden on us to keep them working
-  against a fast-moving core, for a backend we don't own. Ship them as a
-  **standalone plugin repo** users install into `~/.hermes/plugins/` (or via a
-  pip entry point), and promote them in the Nous Research Discord
-  (`#plugins-skills-and-skins`). This is a coupling-and-maintenance decision, not
-  a quality bar — the plugin can be excellent and still be a close. PRs that add
-  such a directory to the tree are closed with a pointer to publish it as its own
-  repo.
 
 ### Before you call it a bug — verify the premise (and when NOT to close)
 
@@ -491,18 +480,18 @@ The dashboard embeds the real `hermes --tui` — **not** a rewrite.  See `hermes
 
 ### Electron Desktop Chat App (`apps/desktop/`)
 
-A **separate** chat surface from both the classic CLI and the dashboard's embedded TUI. It is an Electron + React + nanostore renderer (`@assistant-ui/react`) that talks to a `tui_gateway` backend over JSON-RPC (`requestGateway(method, params)`). The WebSocket/JSON-RPC transport lives in the framework-agnostic `apps/shared` package (`@hermes/shared` — `JsonRpcGatewayClient` + WS URL helpers), which the web dashboard (`web/`) also consumes; **desktop has no build/runtime dependency on the dashboard frontend** — it spawns a headless `hermes serve` backend server (the same gateway `dashboard` serves, minus the browser UI entirely: `serve` sets `headless_backend=True`, so `cmd_dashboard` skips `_build_web_ui` AND exports `HERMES_SERVE_HEADLESS=1` so `mount_spa()` disables the SPA even if a stray `web_dist/` exists — only the JSON-RPC/WS/API surface is reachable). `dashboard` and `serve` share `cmd_dashboard`/`start_server` but are independent surfaces — neither launches the other. The one exception is a backward-compat *fallback*: `serve` is newer, so the desktop spawn (`electron/backend-command.ts` + `backendSupportsServe()` in `electron/main.ts`) detects whether the resolved runtime registers `serve` and, only when it does not (an older managed install / PATH `hermes` the app hasn't updated yet), rewrites the argv to the legacy `dashboard --no-open`. Without that, a new app against an un-upgraded runtime would crash on an unknown subcommand and brick every mid-upgrade user. It does NOT embed `hermes --tui` — it has its own composer, transcript, and slash-command pipeline. For scoped Desktop architecture, state, resolver, transport, and testing rules, read `apps/desktop/AGENTS.md`.
+A **separate** chat surface from both the classic CLI and the dashboard's embedded TUI. It is an Electron + React + nanostore renderer (`@assistant-ui/react`) that talks to a `tui_gateway` backend over JSON-RPC (`requestGateway(method, params)`). It does NOT embed `hermes --tui` — it has its own composer, transcript, and slash-command pipeline. Route desktop bugs to the `hermes-desktop-app-work` skill, not `hermes-dashboard-work`.
 
 **Slash commands in the desktop app are curated client-side, then dispatched to the backend.** The pipeline:
 
 - **Backend already provides everything.** `tui_gateway/server.py` `commands.catalog` (empty-query list) and `complete.slash` (typed-query completions) both include built-in commands, user `quick_commands`, AND skill-derived commands (`scan_skill_commands()` / `get_skill_commands()`). The desktop app does not need a new RPC to see skills.
-- **The renderer curates via `apps/desktop/src/lib/desktop-slash-commands.ts`.** This is the load-bearing file. It holds `DESKTOP_COMMAND_SPECS` (the built-ins and their Desktop surfaces) plus `NO_DESKTOP_SURFACE` block-lists for terminal-only / messaging-only / picker-owned / settings-owned / advanced commands that should NOT clutter the desktop popover.
+- **The renderer curates via `apps/desktop/src/lib/desktop-slash-commands.ts`.** This is the load-bearing file. It holds `DESKTOP_COMMANDS` (the ~19 built-ins shown in the palette) plus block-lists for terminal-only / messaging-only / picker-owned / settings-owned / advanced commands that should NOT clutter the desktop popover.
   - `isDesktopSlashCommand(name)` — gates **execution**. Returns true for built-ins AND for any non-built-in (skill / quick command), so typed extension commands run.
   - `isDesktopSlashSuggestion(name)` — gates **discovery/completion**. Used by BOTH completion paths in `app/chat/composer/hooks/use-slash-completions.ts` (empty-query catalog filter + typed-query `complete.slash` filter) and by `filterDesktopCommandsCatalog`.
   - `isDesktopSlashExtensionCommand(name)` — true when the command is NOT a known Hermes built-in (i.e. a skill or user quick command). Both suggestion and catalog-filter paths allow extensions through so skill commands surface in the palette. (Added when fixing "skill commands missing from the desktop slash palette" — the curated allow-list was silently dropping every skill/quick command from completions even though they executed fine when typed.)
-- **Dispatch** lives in `app/session/hooks/use-prompt-actions/slash.ts` (`runSlash`): built-ins that the desktop owns (`/skin`, `/help`, `/new`, …) are handled locally or via `commands.catalog`; everything else goes to `slash.exec`, falling back to `command.dispatch` (which the gateway resolves into skill / alias / exec directives). A skill command resolves to `{type: "skill", message}` and is submitted as a normal prompt.
+- **Dispatch** lives in `app/session/hooks/use-prompt-actions.ts` (`runSlash`): built-ins that the desktop owns (`/skin`, `/help`, `/new`, …) are handled locally or via `commands.catalog`; everything else goes to `slash.exec`, falling back to `command.dispatch` (which the gateway resolves into skill / alias / exec directives). A skill command resolves to `{type: "skill", message}` and is submitted as a normal prompt.
 
-**Rule:** the desktop slash palette's curation is about hiding noise (terminal-only / messaging-only built-ins), NOT about hiding user-activated extensions. Skill commands and `quick_commands` are extensions the backend surfaces — they belong in completions. If you tighten `desktop-slash-commands.ts`, keep `isDesktopSlashExtensionCommand` flowing into both the suggestion and catalog-filter paths. Tests: from `apps/desktop`, run `npx vitest run src/lib/desktop-slash-commands.test.ts` (workspace dependencies are installed at the repo root).
+**Rule:** the desktop slash palette's curation is about hiding noise (terminal-only / messaging-only built-ins), NOT about hiding user-activated extensions. Skill commands and `quick_commands` are extensions the backend surfaces — they belong in completions. If you tighten `desktop-slash-commands.ts`, keep `isDesktopSlashExtensionCommand` flowing into both the suggestion and catalog-filter paths. Tests: `apps/desktop/src/lib/desktop-slash-commands.test.ts` (run via the repo-root `vitest`, since `apps/desktop` resolves deps from the root workspace install).
 
 ---
 
@@ -794,24 +783,6 @@ landing in this tree. PRs that add a new directory under
 provider as its own repo. Existing in-tree providers stay; bug fixes
 to them are welcome.
 
-**No new third-party-product plugins in-tree (policy, June 2026):** the
-same rule applies beyond memory providers. Plugins that integrate
-someone else's product or project — observability/metrics backends,
-vendor SaaS connectors, analytics dashboards, paid-service tie-ins —
-must ship as **standalone plugin repos** that users install into
-`~/.hermes/plugins/` (or via pip entry points). They register through
-the existing plugin discovery path and use the ABCs/hooks/ctx surface
-we expose; nothing special is needed in core. The reason is
-maintenance load: every product we absorb into the tree becomes our
-burden to keep working against a fast-moving core, for a backend we
-don't own. Promote standalone plugins in the Nous Research Discord
-(`#plugins-skills-and-skins`). PRs that add such a directory under
-`plugins/` are closed with a pointer to publish it as its own repo —
-this is a coupling decision, not a quality judgment. (The
-`observability/`, `kanban/`, `disk-cleanup/`, etc. directories already
-in the tree are existing precedent, not an invitation to add more
-third-party-product plugins alongside them.)
-
 ### Model-provider plugins (`plugins/model-providers/<name>/`)
 
 Every inference backend (openrouter, anthropic, gmi, deepseek, nvidia, …)
@@ -998,8 +969,7 @@ Two shapes:
 Roles:
 
 - `role="leaf"` (default) — focused worker. Cannot call `delegate_task`,
-  `clarify`, `memory`, `send_message`, `cronjob`. Retains `execute_code`
-  (programmatic tool calling).
+  `clarify`, `memory`, `send_message`, `execute_code`.
 - `role="orchestrator"` — retains `delegate_task` so it can spawn its
   own workers. Gated by `delegation.orchestrator_enabled` (default true)
   and bounded by `delegation.max_spawn_depth` (default 2).
@@ -1095,16 +1065,14 @@ kanban task.
 
 - **CLI:** `hermes_cli/kanban.py` wires `hermes kanban` with verbs
   `init`, `create`, `list` (alias `ls`), `show`, `assign`, `link`,
-  `unlink`, `comment`, `attach`, `attachments`, `attach-rm`, `complete`,
-  `block`, `unblock`, `archive`, `tail`, plus less-commonly-used `watch`,
-  `stats`, `runs`, `log`, `assignees`, `heartbeat`, `notify-*`,
-  `dispatch`, `daemon`, `gc`.
+  `unlink`, `comment`, `complete`, `block`, `unblock`, `archive`,
+  `tail`, plus less-commonly-used `watch`, `stats`, `runs`, `log`,
+  `assignees`, `heartbeat`, `notify-*`, `dispatch`, `daemon`, `gc`.
 - **Worker/orchestrator toolset:** `tools/kanban_tools.py` exposes
   `kanban_show`, `kanban_complete`, `kanban_block`, `kanban_heartbeat`,
-  `kanban_comment`, `kanban_create`, `kanban_link`, `kanban_attach`,
-  `kanban_attach_url`, `kanban_attachments`; profiles that explicitly
-  enable the `kanban` toolset outside a dispatcher-spawned task also get
-  `kanban_list` and `kanban_unblock` for board routing.
+  `kanban_comment`, `kanban_create`, `kanban_link`; profiles that
+  explicitly enable the `kanban` toolset outside a dispatcher-spawned
+  task also get `kanban_list` and `kanban_unblock` for board routing.
 - **Dispatcher:** long-lived loop that (default every 60s) reclaims
   stale claims, promotes ready tasks, atomically claims, and spawns
   assigned profiles. Runs **inside the gateway** by default via
@@ -1207,7 +1175,7 @@ automatically scope to the active profile.
    a unique credential (bot token, API key), call `acquire_scoped_lock()` from
    `gateway.status` in the `connect()`/`start()` method and `release_scoped_lock()` in
    `disconnect()`/`stop()`. This prevents two profiles from using the same credential.
-   See `plugins/platforms/irc/adapter.py` for the canonical pattern.
+   See `gateway/platforms/telegram.py` for the canonical pattern.
 
 6. **Profile operations are HOME-anchored, not HERMES_HOME-anchored** — `_get_profiles_root()`
    returns `Path.home() / ".hermes" / "profiles"`, NOT `get_hermes_home() / "profiles"`.
@@ -1281,7 +1249,6 @@ def profile_env(tmp_path, monkeypatch):
 
 ## Testing
 
-### Python
 **ALWAYS use `scripts/run_tests.sh`** — do not call `pytest` directly. The script enforces
 hermetic environment parity with CI (unset credential vars, TZ=UTC, LANG=C.UTF-8,
 `-n auto` xdist workers, in-tree subprocess-isolation plugin). Direct `pytest`
@@ -1293,41 +1260,65 @@ scripts/run_tests.sh                                  # full suite, CI-parity
 scripts/run_tests.sh tests/gateway/                   # one directory
 scripts/run_tests.sh tests/agent/test_foo.py::test_x  # one test
 scripts/run_tests.sh -v --tb=long                     # pass-through pytest flags
+scripts/run_tests.sh --no-isolate tests/foo/          # disable subprocess isolation (faster, for debugging)
 ```
 
-**Flake policy:** the runner auto-retries a failing test FILE once in a fresh
-subprocess (`--file-retries`, default 1; `HERMES_TEST_FILE_RETRIES=0` to
-disable). Pass-on-retry counts as green but is printed in a `⚠ FLAKY` summary
-section with both attempts' output. A FLAKY report is a bug to fix, not noise
-to ignore — timing-sensitive tests must not assume a quiet runner (loose
-wall-clock bounds ≥ 2s, event-based sync, no `assert not _wait_until(...)`
-negative-timing races).
+### Subprocess-per-test isolation
 
-#### Subprocess-per-test-file isolation
+Every test runs in a freshly-spawned Python subprocess via the in-tree plugin
+at `tests/_isolate_plugin.py`. This means module-level dicts/sets and
+ContextVars from one test cannot leak into the next — the historic
+`_reset_module_state` autouse fixture is gone.
 
-Every test file runs in a freshly-spawned Python subprocess via `run_tests_parallel.py`. This means module-level dicts/sets and
-ContextVars from one test file cannot leak into the next.
+Implementation notes:
 
-#### Why the wrapper
+- The plugin uses `multiprocessing.get_context("spawn")`, which works on
+  Linux, macOS, and Windows alike (POSIX `fork` is not used).
+- Per-test overhead is ~0.5–1.0s (Python startup + pytest collection). xdist
+  parallelism amortizes this across cores; on a 20-core box the full suite
+  finishes in roughly the same wall time as before, but flake-free.
+- `isolate_timeout` (configured in `pyproject.toml`) caps each test at 30s.
+  Hangs are killed and surfaced as a failure report.
+- Pass `--no-isolate` to disable isolation — useful when debugging a single
+  test interactively, or when you specifically want to verify state leakage.
+- The plugin disables itself in child processes (sentinel envvar
+  `HERMES_ISOLATE_CHILD=1`), so there's no fork-bomb risk.
 
-|                     | Without wrapper                             | With wrapper                              |
-| ------------------- | ------------------------------------------- | ----------------------------------------- |
-| Provider API keys   | Whatever is in your env (auto-detects pool) | All env vars except a specific few unset. |
-| HOME / `~/.hermes/` | Your real config+auth.json                  | Temp dir per test                         |
-| Timezone            | Local TZ (PDT etc.)                         | UTC                                       |
-| Locale              | Whatever is set                             | C.UTF-8                                   |
+### Why the wrapper (and why the old "just call pytest" doesn't work)
 
-### Where to place what tests
+Five real sources of local-vs-CI drift the script closes:
 
-The CI change classifier (`scripts/ci/classify_changes.py`) runs specific jobs based on what files changed. A Python test that asserts
-about the contents of `package.json`, `package-lock.json`, `.ts`/`.tsx`
-source, or any other JS-side artifact will not run on a PR that only touches
-those files. This means a regression can go green on a PR and red on `main` (where the
-classifier fails open and runs everything).
+| | Without wrapper | With wrapper |
+|---|---|---|
+| Provider API keys | Whatever is in your env (auto-detects pool) | All `*_API_KEY`/`*_TOKEN`/etc. unset |
+| HOME / `~/.hermes/` | Your real config+auth.json | Temp dir per test |
+| Timezone | Local TZ (PDT etc.) | UTC |
+| Locale | Whatever is set | C.UTF-8 |
+| xdist workers | `-n auto` = all cores | `-n auto` (safe — subprocess isolation prevents cross-worker flakes) |
 
-Any test that reads or asserts about `package.json`,
-`package-lock.json`, `tsconfig.json`, `.ts`/`.tsx`/`.js`/`.mjs`/`.cjs`
-source files configuration belongs in the JS (vitest) test suite, not in `tests/*.py`.
+`tests/conftest.py` also enforces points 1-4 as an autouse fixture so ANY pytest
+invocation (including IDE integrations) gets hermetic behavior — but the wrapper
+is belt-and-suspenders.
+
+### Running without the wrapper (only if you must)
+
+If you can't use the wrapper (e.g. inside an IDE that shells pytest directly),
+at minimum activate the venv. The isolation plugin loads automatically from
+`addopts` in `pyproject.toml`, so you get the same per-test process isolation
+either way.
+
+```bash
+source .venv/bin/activate   # or: source venv/bin/activate
+python -m pytest tests/ -q
+```
+
+If you need to bypass isolation for fast feedback while debugging:
+
+```bash
+python -m pytest tests/agent/test_foo.py -q --no-isolate
+```
+
+Always run the full suite before pushing changes.
 
 ### Don't write change-detector tests
 
@@ -1377,58 +1368,3 @@ not the specific names.
 
 Reviewers should reject new change-detector tests; authors should convert
 them into invariants before re-requesting review.
-
-### Never read source code in tests
-
-A test that reads a source file's text is testing *the shape of the
-source code*, not its behavior. This is a hard antipattern, banned outright.
-Any test that reads a .py, .ts, .tsx, etc., file is suspect.
-
-**Why it's actively harmful, not just weak:**
-
-- It passes when the implementation is subtly broken (the regex matches a
-  call site that exists but is wired wrong) and fails when a correct
-  refactor changes formatting, variable names, or control flow with
-  identical runtime behavior. Both directions of failure are wrong.
-- It can't be run against a built/bundled/minified artifact, so it silently
-  stops testing anything the moment code moves, gets renamed, or a
-  dependency reformats it.
-- It actively blocks refactors: reviewers see "keeps a pattern intact" tests
-  fail during pure structural cleanup with no behavior change, and either
-  hand-wave the failure (dangerous) or waste time updating regexes that add
-  nothing (waste).
-- It gives false confidence. a green suite full of source-regex tests
-  looks like coverage but has never once executed the code path it claims
-  to guard.
-
-**Do not write:**
-
-```ts
-const source = fs.readFileSync(path.join(__dirname, 'main.ts'), 'utf8')
-
-test('backend spawn hides the Windows console', () => {
-  assert.match(source, /spawn\(\s*backend\.command,\s*backend\.args[\s\S]{0,300}hiddenWindowsChildOptions/)
-})
-```
-
-**Do write — extract the logic into a small pure/DI-testable function and
-call it for real:**
-
-```ts
-// backend-spawn.ts
-export function hiddenWindowsChildOptions(options: SpawnOptionsLike = {}, isWindows = process.platform === 'win32') {
-  if (!isWindows || 'windowsHide' in options) return options
-  return { ...options, windowsHide: true }
-}
-
-// backend-spawn.test.ts
-test('windowsHide defaults to true on Windows, is left alone elsewhere', () => {
-  assert.equal(hiddenWindowsChildOptions({}, true).windowsHide, true)
-  assert.equal(hiddenWindowsChildOptions({}, false).windowsHide, undefined)
-  assert.equal(hiddenWindowsChildOptions({ windowsHide: false }, true).windowsHide, false)
-})
-```
-
-If the logic lives inline in a god-file (`main.ts`, `cli.py`,
-`gateway/run.py`) and extracting it feels disruptive: that's the actual
-signal to do the extraction, not to regex around it.

CONTRIBUTING.es.md

@@ -74,7 +74,7 @@ Esto no es una barra de calidad — es una decisión de acoplamiento y mantenimi
 | Requisito | Notas |
 |-----------|-------|
 | **Git** | Con la extensión `git-lfs` instalada |
-| **Python 3.11–3.13** | uv lo instalará si falta |
+| **Python 3.11+** | uv lo instalará si falta |
 | **uv** | Gestor de paquetes Python rápido ([instalar](https://docs.astral.sh/uv/)) |
 | **Node.js 20+** | Opcional — necesario para herramientas de navegador y puente WhatsApp (coincide con los engines de `package.json` raíz) |
 

CONTRIBUTING.md

@@ -85,23 +85,6 @@ This isn't a quality bar — it's a coupling-and-maintenance decision. Memory pr
 
 ---
 
-## Third-Party Product Integrations: Ship as a Standalone Plugin
-
-The same rule extends to **any plugin that integrates someone else's product or project** — observability/metrics backends, vendor SaaS connectors, analytics dashboards, paid-service tie-ins, and similar third-party integrations. **These do not land in this repo.**
-
-The reason is maintenance load, not quality. Every external product absorbed into the core tree becomes ours to keep working against a fast-moving codebase, for a backend we don't own and can't control. Hermes ships a lot and the core moves quickly; coupling third-party products into it creates an open-ended burden on the maintainers.
-
-Publish these as a **standalone plugin repo** instead:
-
-- Implement the relevant ABC and use the existing plugin discovery path (`~/.hermes/plugins/`, project `.hermes/plugins/`, or a pip entry point) — see [Build a Hermes Plugin](https://hermes-agent.nousresearch.com/docs/guides/build-a-hermes-plugin)
-- Register lifecycle hooks (`pre_tool_call`, `post_tool_call`, `pre_llm_call`, `post_llm_call`, `on_session_start`, `on_session_end`), tools (`ctx.register_tool`), and CLI subcommands (`ctx.register_cli_command`) through the surface we already expose — no core changes needed
-- If your plugin needs a capability the framework doesn't expose, that's a feature request to **widen the generic plugin surface** (a new hook or `ctx` method) — never special-case your plugin in core
-- Promote it in the [Nous Research Discord](https://discord.gg/NousResearch) `#plugins-skills-and-skins` channel so users can find and install it
-
-A well-built third-party-product plugin can clear automated review and still be closed for this reason — it's a placement decision, not a verdict on the code. PRs that add such a directory under `plugins/` will be closed with a pointer to publish it as its own repo.
-
----
-
 ## Development Setup
 
 ### Prerequisites
@@ -109,7 +92,7 @@ A well-built third-party-product plugin can clear automated review and still be
 | Requirement | Notes |
 |-------------|-------|
 | **Git** | With the `git-lfs` extension installed |
-| **Python 3.11–3.13** | uv will install it if missing |
+| **Python 3.11+** | uv will install it if missing |
 | **uv** | Fast Python package manager ([install](https://docs.astral.sh/uv/)) |
 | **Node.js 20+** | Optional — needed for browser tools and WhatsApp bridge (matches root `package.json` engines) |
 
@@ -149,20 +132,13 @@ this way, make sure you run the `hermes` entrypoint from this venv; running the
 system `python3 -m hermes_cli.main` can pick up unrelated system Python
 packages.
 
-Create the venv **outside** the cloned source tree. A venv that lives inside
-the directory the agent operates from can be wiped by a relative-path command
-the agent runs against its own checkout (`rm -rf venv`, `uv venv venv`, etc.),
-which silently destroys the running runtime mid-session. Keeping it outside the
-tree means no relative path from the workspace resolves to it.
-
 ```bash
 git clone https://github.com/NousResearch/hermes-agent.git
 cd hermes-agent
 
-# Create venv with Python 3.11, OUTSIDE the source tree
-uv venv ~/.hermes/venvs/hermes-dev --python 3.11
-export VIRTUAL_ENV="$HOME/.hermes/venvs/hermes-dev"
-export PATH="$VIRTUAL_ENV/bin:$PATH"
+# Create venv with Python 3.11
+uv venv venv --python 3.11
+export VIRTUAL_ENV="$(pwd)/venv"
 
 # Install with all extras (messaging, cron, CLI menus, dev tools)
 uv pip install -e ".[all,dev]"

Dockerfile

@@ -26,8 +26,8 @@ ENV PLAYWRIGHT_BROWSERS_PATH=/opt/hermes/.playwright
 # replaces tini with s6-overlay's /init (PID 1 = s6-svscan), which reaps
 # zombies non-blockingly on SIGCHLD and additionally supervises the main
 # hermes process, the dashboard, and per-profile gateways.
-RUN apt-get -o Acquire::Retries=3 update && \
-    apt-get -o Acquire::Retries=3 install -y --no-install-recommends \
+RUN apt-get update && \
+    apt-get install -y --no-install-recommends \
     ca-certificates curl iputils-ping python3 python-is-python3 ripgrep ffmpeg gcc g++ make cmake python3-dev python3-venv libffi-dev libolm-dev procps git openssh-client docker-cli xz-utils && \
     rm -rf /var/lib/apt/lists/*
 
@@ -40,30 +40,33 @@ RUN apt-get -o Acquire::Retries=3 update && \
 # we map between them inline. The noarch + symlinks tarballs are
 # architecture-independent and reused as-is.
 #
-# We use `curl` instead of `ADD` for ALL three tarballs: `ADD` evaluates its
-# URL at parse time (no ARG / TARGETARCH substitution) and — critically for
-# CI reliability — cannot retry, so a single GitHub-release CDN blip fails
-# the whole 15-45 min build. curl -fsSL --retry 3 self-heals those blips,
-# and every tarball is still checksum-verified below before extraction.
+# We use `curl` instead of `ADD` for the per-arch tarball because `ADD`
+# evaluates its URL at parse time, before any ARG / TARGETARCH substitution
+# — splitting one URL per arch into two ADDs would download both on every
+# build and leave dead bytes in the cache. A single curl + arch-keyed URL
+# is simpler and cache-friendlier.
+#
+# Supply-chain integrity: every tarball is checksum-verified against the
+# upstream-published SHA256. To bump S6_OVERLAY_VERSION, fetch the four
+# `.sha256` files from the corresponding release and update the ARGs. The
+# checksum lookup happens during build, so a compromised release artifact
+# fails the build loudly instead of silently producing a tampered image.
 ARG TARGETARCH
 ARG S6_OVERLAY_VERSION=3.2.3.0
 ARG S6_OVERLAY_NOARCH_SHA256=b720f9d9340efc8bb07528b9743813c836e4b02f8693d90241f047998b4c53cf
 ARG S6_OVERLAY_X86_64_SHA256=a93f02882c6ed46b21e7adb5c0add86154f01236c93cd82c7d682722e8840563
 ARG S6_OVERLAY_AARCH64_SHA256=0952056ff913482163cc30e35b2e944b507ba1025d78f5becbb89367bf344581
 ARG S6_OVERLAY_SYMLINKS_SHA256=a60dc5235de3ecbcf874b9c1f18d73263ab99b289b9329aa950e8729c4789f0e
+ADD https://github.com/just-containers/s6-overlay/releases/download/v${S6_OVERLAY_VERSION}/s6-overlay-noarch.tar.xz /tmp/
+ADD https://github.com/just-containers/s6-overlay/releases/download/v${S6_OVERLAY_VERSION}/s6-overlay-symlinks-noarch.tar.xz /tmp/
 RUN set -eu; \
     case "${TARGETARCH:-amd64}" in \
         amd64) s6_arch="x86_64"; s6_arch_sha="${S6_OVERLAY_X86_64_SHA256}" ;; \
         arm64) s6_arch="aarch64"; s6_arch_sha="${S6_OVERLAY_AARCH64_SHA256}" ;; \
         *) echo "Unsupported TARGETARCH=${TARGETARCH} for s6-overlay" >&2; exit 1 ;; \
     esac; \
-    base="https://github.com/just-containers/s6-overlay/releases/download/v${S6_OVERLAY_VERSION}"; \
-    curl -fsSL --retry 3 -o /tmp/s6-overlay-noarch.tar.xz \
-        "${base}/s6-overlay-noarch.tar.xz"; \
-    curl -fsSL --retry 3 -o /tmp/s6-overlay-symlinks-noarch.tar.xz \
-        "${base}/s6-overlay-symlinks-noarch.tar.xz"; \
     curl -fsSL --retry 3 -o /tmp/s6-overlay-arch.tar.xz \
-        "${base}/s6-overlay-${s6_arch}.tar.xz"; \
+        "https://github.com/just-containers/s6-overlay/releases/download/v${S6_OVERLAY_VERSION}/s6-overlay-${s6_arch}.tar.xz"; \
     { \
         printf '%s  %s\n' "${S6_OVERLAY_NOARCH_SHA256}" /tmp/s6-overlay-noarch.tar.xz; \
         printf '%s  %s\n' "${s6_arch_sha}" /tmp/s6-overlay-arch.tar.xz; \
@@ -73,19 +76,17 @@ RUN set -eu; \
     tar -C / -Jxpf /tmp/s6-overlay-noarch.tar.xz; \
     tar -C / -Jxpf /tmp/s6-overlay-arch.tar.xz; \
     tar -C / -Jxpf /tmp/s6-overlay-symlinks-noarch.tar.xz; \
-    rm /tmp/s6-overlay-*.tar.xz /tmp/s6-overlay.sha256
-
-# #34192 / #66679: backward-compat shim for orchestration templates that
-# still reference the legacy /usr/bin/tini entrypoint (Hostinger's
-# 'Hermes WebUI' catalog, NAS compose projects that preserve an old
-# entrypoint on image update, etc.). A plain symlink to /init made the
-# path exist, but forwarded tini flags like `-g` into s6-overlay's
-# rc.init as the container CMD (`rc.init: 91: -g: not found`) and
-# boot-looped any `restart: unless-stopped` deploy. The shim strips the
-# tini CLI surface, then exec's /init + main-wrapper — see
-# docker/tini-shim.sh. Safe to drop once the affected catalogs are
-# updated.
-COPY --chmod=0755 docker/tini-shim.sh /usr/bin/tini
+    rm /tmp/s6-overlay-*.tar.xz /tmp/s6-overlay.sha256; \
+    # #34192: backward-compat shim for orchestration templates that still\
+    # reference the legacy /usr/bin/tini entrypoint (e.g. Hostinger's\
+    # 'Hermes WebUI' catalog). The image has moved to s6-overlay /init\
+    # as PID 1 (see ENTRYPOINT below + the migration comment at the top\
+    # of this file), but external wrappers pinned to /usr/bin/tini will\
+    # crash with 'tini: No such file or directory' on startup. The shim\
+    # symlinks /usr/bin/tini -> /init so legacy wrappers exec the right\
+    # PID-1 reaper without behavior change for users on the current\
+    # ENTRYPOINT. Safe to drop once the affected catalogs are updated.\
+    ln -sf /init /usr/bin/tini
 
 # Non-root user for runtime; UID can be overridden via HERMES_UID at runtime
 RUN useradd -u 10000 -m -d /opt/data hermes
@@ -118,9 +119,6 @@ COPY package.json package-lock.json ./
 COPY web/package.json web/
 COPY ui-tui/package.json ui-tui/
 COPY ui-tui/packages/hermes-ink/ ui-tui/packages/hermes-ink/
-# apps/shared/ is copied IN FULL because web/package.json references it as a
-# `file:` workspace dependency (same pattern as hermes-ink above).
-COPY apps/shared/ apps/shared/
 
 # `npm_config_install_links=false` forces npm to install `file:` deps as
 # symlinks instead of copies.  This is the default since npm 10+, which is
@@ -134,11 +132,8 @@ COPY apps/shared/ apps/shared/
 # guards against a future regression if the source npm version changes.
 ENV npm_config_install_links=false
 
-RUN npm install --prefer-offline --no-audit --fetch-retries=5 && \
-    for i in 1 2 3; do \
-        npx playwright install --with-deps chromium --only-shell && break || \
-        { [ "$i" = 3 ] && exit 1; echo "playwright install failed (attempt $i); retrying in 10s"; sleep 10; }; \
-    done && \
+RUN npm install --prefer-offline --no-audit && \
+    npx playwright install --with-deps chromium --only-shell && \
     npm cache clean --force
 
 # ---------- Layer-cached Python dependency install ----------
@@ -189,19 +184,12 @@ RUN uv sync --frozen --no-install-project --extra all --extra messaging --extra
 # invalidate the (relatively slow) web + ui-tui build layer.
 COPY web/ web/
 COPY ui-tui/ ui-tui/
-COPY apps/shared/ apps/shared/
 RUN cd web && npm run build && \
     cd ../ui-tui && npm run build
 
 # ---------- Source code ----------
 # .dockerignore excludes node_modules, so the installs above survive.
-# --link decouples this layer from parents for cache purposes; --chmod bakes
-# the final read-only permissions at copy time so we skip the separate
-# `chmod -R` pass that previously walked ~30k files across the venv +
-# node_modules + source (21s amd64 / 222s arm64 — #49113).  `a+rX,go-w`
-# gives the non-root hermes user read + traverse but no write; root retains
-# write so the build steps below don't need chmod u+w dances.
-COPY --link --chmod=a+rX,go-w . .
+COPY . .
 
 # ---------- Permissions ----------
 # Link hermes-agent itself (editable). Deps are already installed in the
@@ -209,15 +197,19 @@ COPY --link --chmod=a+rX,go-w . .
 # resolution or downloads.
 RUN uv pip install --no-cache-dir --no-deps -e "."
 
-# Wire the exec shim and install-method stamp.  Files under /opt/hermes are
-# already root-owned (COPY, uv sync, npm install all run as root) and
-# read-only for the hermes user (go-w from the --chmod above).
-
+# Keep /opt/hermes immutable for the runtime hermes user. Hosted/container
+# instances must not be able to self-edit the installed source or venv; user
+# data, skills, plugins, config, logs, and dashboard uploads live under
+# /opt/data instead. Root can still repair the image during build/boot, but
+# supervised Hermes processes drop to the non-root hermes user.
 USER root
 RUN mkdir -p /opt/hermes/bin && \
     cp /opt/hermes/docker/hermes-exec-shim.sh /opt/hermes/bin/hermes && \
     chmod 0755 /opt/hermes/bin/hermes && \
-    printf 'docker\n' > /opt/hermes/.install_method
+    printf 'docker\n' > /opt/hermes/.install_method && \
+    chown -R root:root /opt/hermes && \
+    chmod -R a+rX /opt/hermes && \
+    chmod -R a-w /opt/hermes
 # The ``.install_method`` stamp is baked next to the running code (the install
 # tree), NOT into $HERMES_HOME. $HERMES_HOME (/opt/data) is a shared data
 # volume that is commonly bind-mounted from the host and even shared with a
@@ -244,11 +236,13 @@ RUN mkdir -p /opt/hermes/bin && \
 #
 # The arg is optional — local `docker build` without --build-arg simply
 # omits the file, and the runtime falls back to live-git lookup.  CI
-# (.github/workflows/docker.yml) passes ${{ github.sha }} so
+# (.github/workflows/docker-publish.yml) passes ${{ github.sha }} so
 # every published image has it.
 ARG HERMES_GIT_SHA=
 RUN if [ -n "${HERMES_GIT_SHA}" ]; then \
-        printf '%s\n' "${HERMES_GIT_SHA}" > /opt/hermes/.hermes_build_sha; \
+        chmod u+w /opt/hermes && \
+        printf '%s\n' "${HERMES_GIT_SHA}" > /opt/hermes/.hermes_build_sha && \
+        chmod a-w /opt/hermes /opt/hermes/.hermes_build_sha; \
     fi
 
 # ---------- s6-overlay service wiring ----------
@@ -296,19 +290,6 @@ ENV HERMES_TUI_DIR=/opt/hermes/ui-tui
 ENV HERMES_HOME=/opt/data
 ENV HERMES_WRITE_SAFE_ROOT=/opt/data
 ENV HERMES_DISABLE_LAZY_INSTALLS=1
-# The published image seals /opt/hermes (root-owned, read-only) so a runtime
-# lazy install can't mutate the agent's own venv and brick it. But opt-in
-# backends (Firecrawl web search, Exa, Feishu, …) keep their SDKs in
-# tools/lazy_deps.py — deliberately NOT baked into [all] (see pyproject.toml
-# policy 2026-05-12: one quarantined release must not break every install).
-# Redirect those lazy installs to a writable dir on the durable data volume.
-# lazy_deps appends this dir to the END of sys.path, so a package installed
-# here can only ADD modules — it can never shadow or downgrade a core module,
-# so the sealed-venv guarantee holds even with installs re-enabled. The dir
-# is seeded + chowned to the hermes user by docker/stage2-hook.sh and lives
-# on the /opt/data volume, so it persists across container recreates / image
-# updates (an ABI stamp invalidates it if a rebuild bumps the interpreter).
-ENV HERMES_LAZY_INSTALL_TARGET=/opt/data/lazy-packages
 
 # `docker exec` privilege-drop shim. When operators run
 # `docker exec <c> hermes ...` they default to root, and any file the

MANIFEST.in

@@ -0,0 +1,11 @@
+graft skills
+graft optional-skills
+graft optional-mcps
+graft locales
+# Bundled plugin manifests (plugin.yaml / plugin.yml). Without these the
+# PluginManager scan (hermes_cli/plugins.py) finds zero plugins on installs
+# built from the sdist (e.g. Homebrew, downstream packagers). package-data
+# below covers the wheel; this covers the sdist. See #34034 / #28149.
+recursive-include plugins plugin.yaml plugin.yml
+global-exclude __pycache__
+global-exclude *.py[cod]

README.md

@@ -18,7 +18,7 @@
 
 **The self-improving AI agent built by [Nous Research](https://nousresearch.com).** It's the only agent with a built-in learning loop — it creates skills from experience, improves them during use, nudges itself to persist knowledge, searches its own past conversations, and builds a deepening model of who you are across sessions. Run it on a $5 VPS, a GPU cluster, or serverless infrastructure that costs nearly nothing when idle. It's not tied to your laptop — talk to it from Telegram while it works on a cloud VM.
 
-Use any model you want — [Nous Portal](https://portal.nousresearch.com), OpenRouter, OpenAI, your own endpoint, and [many others](https://hermes-agent.nousresearch.com/docs/integrations/providers). Switch with `hermes model` — no code changes, no lock-in.
+Use any model you want — [Nous Portal](https://portal.nousresearch.com), [OpenRouter](https://openrouter.ai) (200+ models), [NovitaAI](https://novita.ai) (AI-native cloud for Model API, Agent Sandbox, and GPU Cloud), [NVIDIA NIM](https://build.nvidia.com) (Nemotron), [Xiaomi MiMo](https://platform.xiaomimimo.com), [z.ai/GLM](https://z.ai), [Kimi/Moonshot](https://platform.moonshot.ai), [MiniMax](https://www.minimax.io), [Hugging Face](https://huggingface.co), OpenAI, or your own endpoint. Switch with `hermes model` — no code changes, no lock-in.
 
 <table>
 <tr><td><b>A real terminal interface</b></td><td>Full TUI with multiline editing, slash-command autocomplete, conversation history, interrupt-and-redirect, and streaming tool output.</td></tr>
@@ -109,7 +109,6 @@ hermes              # Interactive CLI — start a conversation
 hermes model        # Choose your LLM provider and model
 hermes tools        # Configure which tools are enabled
 hermes config set   # Set individual config values
-hermes config get   # Print individual config values
 hermes gateway      # Start the messaging gateway (Telegram, Discord, etc.)
 hermes setup        # Run the full setup wizard (configures everything at once)
 hermes claw migrate # Migrate from OpenClaw (if coming from OpenClaw)
@@ -233,14 +232,10 @@ scripts/run_tests.sh
 Manual clone fallback (for throwaway clones/CI where you intentionally do not
 want the managed install layout):
 
-Create the venv outside the cloned source tree — a venv inside the directory
-the agent operates from can be wiped by a relative-path command the agent runs
-against its own checkout, destroying the running runtime mid-session.
-
 ```bash
 curl -LsSf https://astral.sh/uv/install.sh | sh
-uv venv ~/.hermes/venvs/hermes-dev --python 3.11
-source ~/.hermes/venvs/hermes-dev/bin/activate
+uv venv .venv --python 3.11
+source .venv/bin/activate
 uv pip install -e ".[all,dev]"
 scripts/run_tests.sh
 ```

acp_adapter/edit_approval.py

@@ -10,7 +10,6 @@ from __future__ import annotations
 import asyncio
 import json
 import logging
-import re
 import tempfile
 from concurrent.futures import TimeoutError as FutureTimeout
 from contextvars import ContextVar, Token
@@ -128,64 +127,13 @@ def _proposal_for_patch_replace(arguments: dict[str, Any]) -> EditProposal:
     )
 
 
-def _extract_v4a_patch_paths(patch_body: str) -> list[str]:
-    paths: list[str] = []
-    for match in re.finditer(
-        r'^\*\*\*\s+(?:Update|Add|Delete)\s+File:\s*(.+)$',
-        patch_body,
-        re.MULTILINE,
-    ):
-        path = match.group(1).strip()
-        if path:
-            paths.append(path)
-    for match in re.finditer(
-        r'^\*\*\*\s+Move\s+File:\s*(.+?)\s*->\s*(.+)$',
-        patch_body,
-        re.MULTILINE,
-    ):
-        src = match.group(1).strip()
-        dst = match.group(2).strip()
-        if src:
-            paths.append(src)
-        if dst:
-            paths.append(dst)
-    return paths
-
-
-def _proposal_for_patch_v4a(arguments: dict[str, Any]) -> EditProposal:
-    patch_body = arguments.get("patch")
-    if not isinstance(patch_body, str) or not patch_body:
-        raise ValueError("patch content required")
-
-    paths = _extract_v4a_patch_paths(patch_body)
-    if not paths:
-        raise ValueError("no file paths found in V4A patch")
-
-    proposal_path = paths[0] if len(paths) == 1 else ", ".join(paths)
-    old_text = _read_text_if_exists(paths[0]) if len(paths) == 1 else None
-    return EditProposal(
-        tool_name="patch",
-        path=proposal_path,
-        old_text=old_text,
-        # ACP only supports a single diff payload here.  Surface the exact V4A
-        # patch content before execution so patch-mode calls are permissioned
-        # and denied patches cannot mutate.
-        new_text=patch_body,
-        arguments=dict(arguments),
-    )
-
-
 def build_edit_proposal(tool_name: str, arguments: dict[str, Any]) -> EditProposal | None:
     """Return an edit proposal for supported file mutation calls."""
 
     if tool_name == "write_file":
         return _proposal_for_write_file(arguments)
-    if tool_name == "patch":
-        mode = arguments.get("mode", "replace")
-        if mode == "replace":
-            return _proposal_for_patch_replace(arguments)
-        if mode == "patch":
-            return _proposal_for_patch_v4a(arguments)
+    if tool_name == "patch" and arguments.get("mode", "replace") == "replace":
+        return _proposal_for_patch_replace(arguments)
     return None
 
 

acp_adapter/entry.py

@@ -23,16 +23,10 @@ except ModuleNotFoundError:
     # new code but ``uv pip install -e .`` didn't finish.  Missing bootstrap
     # means UTF-8 stdio setup is skipped on Windows; POSIX is unaffected.
     pass
-else:
-    # Stop a ``utils/``/``proxy/``/``ui/`` package in the launch directory from
-    # shadowing Hermes's own modules — ``hermes acp`` can be started from any
-    # cwd, including a project that has same-named packages on its path.
-    hermes_bootstrap.harden_import_path()
 
 import argparse
 import asyncio
 import logging
-import os
 import sys
 from pathlib import Path
 from hermes_constants import get_hermes_home
@@ -191,7 +185,7 @@ def _run_setup_browser(assume_yes: bool = False) -> int:
     """Bootstrap agent-browser + Chromium.
 
     Routes through dep_ensure -> install.{sh,ps1} --ensure, sharing code
-    with the runtime lazy installer.
+    with ``hermes postinstall`` and the runtime lazy installer.
 
     Returns 0 on success, 1 on failure.
     """
@@ -252,13 +246,11 @@ def main(argv: list[str] | None = None) -> None:
     # MCP servers dynamically via asyncio.to_thread inside the event
     # loop; that path is unaffected.)  Moved from model_tools.py module
     # scope to avoid freezing the gateway's loop on lazy import (#16856).
-    # Metadata-only hosts can opt out of unrelated global MCP startup.
-    if os.environ.get("HERMES_ACP_SKIP_CONFIGURED_MCP", "").strip() != "1":
-        try:
-            from tools.mcp_tool import discover_mcp_tools
-            discover_mcp_tools()
-        except Exception:
-            logger.debug("MCP tool discovery failed at ACP startup", exc_info=True)
+    try:
+        from tools.mcp_tool import discover_mcp_tools
+        discover_mcp_tools()
+    except Exception:
+        logger.debug("MCP tool discovery failed at ACP startup", exc_info=True)
 
     agent = HermesACPAgent()
     try:

acp_adapter/permissions.py

@@ -38,22 +38,19 @@ def _permission_option_supports_kind(kind: str) -> bool:
     return True
 
 
-def _build_permission_options(
-    *, allow_permanent: bool, smart_denied: bool = False,
-) -> list[PermissionOption]:
+def _build_permission_options(*, allow_permanent: bool) -> list[PermissionOption]:
     """Return ACP options that match Hermes approval semantics."""
-    options = [PermissionOption(
-        option_id="allow_once", kind="allow_once", name="Allow once",
-    )]
-    if not smart_denied:
-        options.append(PermissionOption(
+    options = [
+        PermissionOption(option_id="allow_once", kind="allow_once", name="Allow once"),
+        PermissionOption(
             option_id="allow_session",
             # ACP has no session-scoped kind, so use the closest persistent
             # hint while keeping Hermes semantics in the option id.
             kind="allow_always",
             name="Allow for session",
-        ))
-    if allow_permanent and not smart_denied:
+        ),
+    ]
+    if allow_permanent:
         options.append(
             PermissionOption(
                 option_id="allow_always",
@@ -62,7 +59,7 @@ def _build_permission_options(
             ),
         )
     options.append(PermissionOption(option_id="deny", kind="reject_once", name="Deny"))
-    if not smart_denied and _permission_option_supports_kind("reject_always"):
+    if _permission_option_supports_kind("reject_always"):
         options.append(
             PermissionOption(
                 option_id="deny_always",
@@ -132,15 +129,11 @@ def make_approval_callback(
         description: str,
         *,
         allow_permanent: bool = True,
-        smart_denied: bool = False,
         **_: object,
     ) -> str:
         from agent.async_utils import safe_schedule_threadsafe
 
-        options = _build_permission_options(
-            allow_permanent=allow_permanent,
-            smart_denied=smart_denied,
-        )
+        options = _build_permission_options(allow_permanent=allow_permanent)
 
         tool_call = _build_permission_tool_call(command, description)
         coro = request_permission_fn(

acp_adapter/server.py

@@ -74,121 +74,9 @@ from acp_adapter.permissions import make_approval_callback
 from acp_adapter.provenance import session_provenance_meta
 from acp_adapter.session import SessionManager, SessionState, _expand_acp_enabled_toolsets
 from acp_adapter.tools import build_tool_complete, build_tool_start
-from agent.context_compressor import (
-    COMPRESSED_SUMMARY_METADATA_KEY,
-    ContextCompressor,
-)
-from tools.approval import (
-    reset_hermes_interactive_context,
-    set_hermes_interactive_context,
-)
 
 logger = logging.getLogger(__name__)
 
-
-def _named_custom_provider_catalogs() -> list[tuple[str, str, list[tuple[str, str]]]]:
-    """Return ``(slug, label, [(model_id, description), ...])`` for named endpoints.
-
-    Covers both the v12 ``providers:`` mapping and the legacy
-    ``custom_providers:`` list.  These endpoints never appear in canonical
-    provider enumeration, so without this the ACP model selector hides every
-    named endpoint that the TUI ``/model`` picker already renders (#47039
-    implemented named-endpoint rows for the TUI surface only).
-
-    Model lists come from the entry's declared models (``default_model`` +
-    ``models``), refreshed from the endpoint's live ``/models`` listing when a
-    credential is available and ``discover_models`` is not disabled.  Declared
-    models are kept even when live discovery fails — some OpenAI-compatible
-    endpoints (e.g. Bedrock Mantle Responses) expose no ``/models`` route at
-    all yet serve the declared models fine.
-
-    Slugs use the ``custom:<name>`` shape that ``parse_model_input`` and
-    ``resolve_runtime_provider`` already resolve, so encoded choice ids
-    (``custom:<name>:<model>``) round-trip through ``set_session_model``
-    unchanged.
-    """
-    try:
-        from hermes_cli.config import (
-            get_compatible_custom_providers,
-            is_provider_enabled,
-            load_config,
-        )
-        from hermes_cli.models import fetch_api_models
-    except ImportError:
-        return []
-
-    try:
-        cfg = load_config()
-        entries = get_compatible_custom_providers(cfg)
-    except Exception:
-        logger.debug("Could not load named custom providers", exc_info=True)
-        return []
-
-    # ``get_compatible_custom_providers`` drops the ``enabled`` flag during
-    # normalization, so collect explicitly disabled provider keys from the
-    # raw config and skip their entries below.
-    disabled_keys: set[str] = set()
-    raw_providers = cfg.get("providers") if isinstance(cfg, dict) else None
-    if isinstance(raw_providers, dict):
-        for raw_key, raw_entry in raw_providers.items():
-            if isinstance(raw_entry, dict) and not is_provider_enabled(raw_entry):
-                disabled_keys.add(str(raw_key).strip().lower())
-
-    catalogs: list[tuple[str, str, list[tuple[str, str]]]] = []
-    for entry in entries:
-        if not isinstance(entry, dict):
-            continue
-        provider_key = str(entry.get("provider_key", "") or "").strip()
-        if provider_key.lower() in disabled_keys:
-            continue
-        name = str(entry.get("name", "") or "").strip()
-        base_url = str(entry.get("base_url", "") or "").strip()
-        if not name or not base_url:
-            continue
-        slug_source = provider_key or name
-        slug = "custom:" + slug_source.strip().lower().replace(" ", "-")
-
-        api_key = str(entry.get("api_key", "") or "").strip()
-        if not api_key:
-            key_env = str(entry.get("key_env", "") or "").strip()
-            api_key = os.environ.get(key_env, "").strip() if key_env else ""
-
-        declared: list[str] = []
-        default_model = str(entry.get("model", "") or "").strip()
-        if default_model:
-            declared.append(default_model)
-        models_cfg = entry.get("models")
-        if isinstance(models_cfg, dict):
-            for mid in models_cfg:
-                mid = str(mid or "").strip()
-                if mid and mid not in declared:
-                    declared.append(mid)
-
-        if not api_key and not declared:
-            # No credential to discover with and nothing declared:
-            # not addressable from the selector.
-            continue
-
-        model_ids = list(declared)
-        discover = entry.get("discover_models", True)
-        if isinstance(discover, str):
-            discover = discover.lower() not in {"false", "no", "0"}
-        if discover and api_key:
-            try:
-                live = fetch_api_models(
-                    api_key, base_url, api_mode=entry.get("api_mode")
-                )
-            except Exception:
-                live = None
-            if live:
-                model_ids = declared + [m for m in live if m not in declared]
-
-        if not model_ids:
-            continue
-        catalogs.append((slug, name, [(mid, "") for mid in model_ids]))
-
-    return catalogs
-
 try:
     from hermes_cli import __version__ as HERMES_VERSION
 except Exception:
@@ -201,13 +89,6 @@ _executor = ThreadPoolExecutor(max_workers=4, thread_name_prefix="acp-agent")
 # does not expose a client-side limit, so this is a fixed cap that clients
 # paginate against using `cursor` / `next_cursor`.
 _LIST_SESSIONS_PAGE_SIZE = 50
-# Per-provider cap for the ACP model selector. ACP clients (Zed, Buzz) render
-# the whole `availableModels` array in one dropdown, so an unbounded
-# cross-provider catalog degrades the picker. Mirrors the cap the MoA picker
-# already uses (`hermes_cli/moa_cmd.py`). This bounds each provider's row, not
-# the total; aggregator providers stay intentionally uncapped inside the shared
-# inventory, and the current model is always kept via the fallback insert below.
-ACP_MAX_MODELS_PER_PROVIDER = 200
 _MAX_ACP_RESOURCE_BYTES = 512 * 1024
 _TEXT_RESOURCE_MIME_PREFIXES = ("text/",)
 _TEXT_RESOURCE_MIME_TYPES = {
@@ -571,7 +452,7 @@ class HermesACPAgent(acp.Agent):
         "tools": "List available tools",
         "context": "Show conversation context info",
         "reset": "Clear conversation history",
-        "compress": "Compress conversation context",
+        "compact": "Compress conversation context",
         "steer": "Inject guidance into the currently running agent turn",
         "queue": "Queue a prompt to run after the current turn finishes",
         "version": "Show Hermes version",
@@ -600,7 +481,7 @@ class HermesACPAgent(acp.Agent):
             "description": "Clear conversation history",
         },
         {
-            "name": "compress",
+            "name": "compact",
             "description": "Compress conversation context",
         },
         {
@@ -696,108 +577,46 @@ class HermesACPAgent(acp.Agent):
         return f"{raw_provider}:{raw_model}"
 
     def _build_model_state(self, state: SessionState) -> SessionModelState | None:
-        """Return authenticated providers and their models for ACP clients.
-
-        The shared Hermes inventory is also used by ``hermes model``, the TUI,
-        and the dashboard. Keeping ACP on that substrate prevents its selector
-        from silently collapsing to the current provider's curated list.
-        """
+        """Return the ACP model selector payload for editors like Zed."""
         model = str(state.model or getattr(state.agent, "model", "") or "").strip()
         provider = getattr(state.agent, "provider", None) or detect_provider() or "openrouter"
 
         try:
-            from hermes_cli.inventory import build_models_payload, load_picker_context
-            from hermes_cli.models import normalize_provider, provider_label
+            from hermes_cli.models import curated_models_for_provider, normalize_provider, provider_label
 
             normalized_provider = normalize_provider(provider)
-            context = load_picker_context().with_overrides(
-                current_provider=normalized_provider,
-                current_model=model,
-                current_base_url=str(getattr(state.agent, "base_url", "") or ""),
-            )
-            payload = build_models_payload(
-                context,
-                explicit_only=True,
-                include_unconfigured=False,
-                picker_hints=False,
-                canonical_order=True,
-                pricing=False,
-                capabilities=False,
-                refresh=False,
-                probe_custom_providers=False,
-                probe_current_custom_provider=False,
-                max_models=ACP_MAX_MODELS_PER_PROVIDER,
-            )
-
+            provider_name = provider_label(normalized_provider)
             available_models: list[ModelInfo] = []
             seen_ids: set[str] = set()
-            for row in payload.get("providers") or []:
-                row_provider = normalize_provider(str(row.get("slug") or "").strip())
-                if not row_provider:
-                    continue
-                provider_name = str(row.get("name") or "").strip() or provider_label(
-                    row_provider
-                )
-                for model_entry in row.get("models") or []:
-                    if isinstance(model_entry, dict):
-                        rendered_model = str(
-                            model_entry.get("id")
-                            or model_entry.get("model")
-                            or model_entry.get("name")
-                            or ""
-                        ).strip()
-                    else:
-                        rendered_model = str(model_entry or "").strip()
-                    if not rendered_model:
-                        continue
-                    choice_id = self._encode_model_choice(row_provider, rendered_model)
-                    if choice_id in seen_ids:
-                        continue
-                    is_current = (
-                        row_provider == normalized_provider and rendered_model == model
-                    )
-                    description = f"Provider: {provider_name}"
-                    if is_current:
-                        description += " • current"
-                    available_models.append(
-                        ModelInfo(
-                            model_id=choice_id,
-                            name=f"{provider_name} · {rendered_model}",
-                            description=description,
-                        )
-                    )
-                    seen_ids.add(choice_id)
 
-            # Named user-defined endpoints (providers: / custom_providers:)
-            # are invisible to canonical provider enumeration — append them
-            # so editor clients can select them like the TUI /model picker.
-            for named_slug, named_label, named_catalog in _named_custom_provider_catalogs():
-                for named_model, named_desc in named_catalog:
-                    named_choice = self._encode_model_choice(named_slug, named_model)
-                    if not named_choice or named_choice in seen_ids:
-                        continue
-                    named_parts = [f"Provider: {named_label}"]
-                    if named_desc:
-                        named_parts.append(str(named_desc).strip())
-                    if named_slug == normalized_provider and named_model == model:
-                        named_parts.append("current")
-                    available_models.append(
-                        ModelInfo(
-                            model_id=named_choice,
-                            name=named_model,
-                            description=" • ".join(part for part in named_parts if part),
-                        )
+            for model_id, description in curated_models_for_provider(normalized_provider):
+                rendered_model = str(model_id or "").strip()
+                if not rendered_model:
+                    continue
+                choice_id = self._encode_model_choice(normalized_provider, rendered_model)
+                if choice_id in seen_ids:
+                    continue
+                desc_parts = [f"Provider: {provider_name}"]
+                if description:
+                    desc_parts.append(str(description).strip())
+                if rendered_model == model:
+                    desc_parts.append("current")
+                available_models.append(
+                    ModelInfo(
+                        model_id=choice_id,
+                        name=rendered_model,
+                        description=" • ".join(part for part in desc_parts if part),
                     )
-                    seen_ids.add(named_choice)
+                )
+                seen_ids.add(choice_id)
 
             current_model_id = self._encode_model_choice(normalized_provider, model)
             if current_model_id and current_model_id not in seen_ids:
-                provider_name = provider_label(normalized_provider)
                 available_models.insert(
                     0,
                     ModelInfo(
                         model_id=current_model_id,
-                        name=f"{provider_name} · {model}",
+                        name=model,
                         description=f"Provider: {provider_name} • current",
                     ),
                 )
@@ -1146,49 +965,11 @@ class HermesACPAgent(acp.Agent):
                 return text
         return ""
 
-    @staticmethod
-    def _history_summary_meta(message: dict[str, Any], text: str) -> dict[str, Any] | None:
-        """Build the ``_meta`` payload for a replayed compaction summary.
-
-        Compaction summaries are persisted as ordinary history messages —
-        standalone handoffs under ``role="user"`` OR ``role="assistant"``
-        (the compressor picks whichever role keeps alternation valid), and
-        merge-into-tail messages where the summary is appended after the
-        first preserved tail message's real content. Without a wire flag,
-        ACP frontends render all of these as ordinary turns.
-
-        Two distinct keys under ``_meta.hermes`` (ACP's extensibility
-        channel), so clients cannot accidentally hide real content:
-
-        * ``compactionSummary: true`` — the entire chunk is the handoff
-          summary. Safe to restyle or collapse wholesale.
-        * ``containsCompactionSummary: true`` — a merged-tail message: real
-          preserved turn content followed by the summary. Clients may style
-          it, but collapsing the whole chunk would hide the preserved
-          content, hence the separate key.
-
-        Detection honors the in-process ``_compressed_summary`` flag and
-        falls back to content classification, so it also works for a
-        DB-reloaded session that lost the in-memory flag.
-        """
-        kind = ContextCompressor.classify_summary_content(text)
-        if kind is None and message.get(COMPRESSED_SUMMARY_METADATA_KEY):
-            # Flagged in-process but content didn't classify (e.g. future
-            # prefix drift): treat as a standalone summary — the flag is only
-            # ever set on summary-bearing messages.
-            kind = "standalone"
-        if kind == "standalone":
-            return {"hermes": {"compactionSummary": True}}
-        if kind == "merged":
-            return {"hermes": {"containsCompactionSummary": True}}
-        return None
-
     @staticmethod
     def _history_message_update(
         *,
         role: str,
         text: str,
-        field_meta: dict[str, Any] | None = None,
     ) -> UserMessageChunk | AgentMessageChunk | None:
         """Build an ACP history replay update for a user/assistant message."""
         block = TextContentBlock(type="text", text=text)
@@ -1196,13 +977,11 @@ class HermesACPAgent(acp.Agent):
             return UserMessageChunk(
                 session_update="user_message_chunk",
                 content=block,
-                field_meta=field_meta,
             )
         if role == "assistant":
             return AgentMessageChunk(
                 session_update="agent_message_chunk",
                 content=block,
-                field_meta=field_meta,
             )
         return None
 
@@ -1273,11 +1052,7 @@ class HermesACPAgent(acp.Agent):
             if role == "user":
                 text = self._history_message_text(message)
                 if text:
-                    update = self._history_message_update(
-                        role=role,
-                        text=text,
-                        field_meta=self._history_summary_meta(message, text),
-                    )
+                    update = self._history_message_update(role=role, text=text)
                     if update is not None and not await _send(update):
                         return
                 continue
@@ -1289,11 +1064,7 @@ class HermesACPAgent(acp.Agent):
 
                 text = self._history_message_text(message)
                 if text:
-                    update = self._history_message_update(
-                        role=role,
-                        text=text,
-                        field_meta=self._history_summary_meta(message, text),
-                    )
+                    update = self._history_message_update(role=role, text=text)
                     if update is not None and not await _send(update):
                         return
 
@@ -1443,19 +1214,12 @@ class HermesACPAgent(acp.Agent):
             with state.runtime_lock:
                 if state.is_running and state.current_prompt_text:
                     state.interrupted_prompt_text = state.current_prompt_text
-                # Publish cancellation and hard-stop the agent before another
-                # prompt can acquire this lock and mistake the turn for
-                # redirectable work.
-                state.cancel_event.set()
-                try:
-                    if getattr(state, "agent", None) and hasattr(state.agent, "interrupt"):
-                        state.agent.interrupt()
-                except Exception:
-                    logger.debug(
-                        "Failed to interrupt ACP session %s",
-                        session_id,
-                        exc_info=True,
-                    )
+            state.cancel_event.set()
+            try:
+                if getattr(state, "agent", None) and hasattr(state.agent, "interrupt"):
+                    state.agent.interrupt()
+            except Exception:
+                logger.debug("Failed to interrupt ACP session %s", session_id, exc_info=True)
             logger.info("Cancelled session %s", session_id)
 
     async def fork_session(
@@ -1584,26 +1348,6 @@ class HermesACPAgent(acp.Agent):
             elif rewrite_idle:
                 user_text = steer_text
                 user_content = steer_text
-        elif (
-            text_only_prompt
-            and isinstance(user_content, str)
-            and not user_text.startswith("/")
-        ):
-            # Some ACP clients implement "stop and send" as two protocol calls:
-            # cancel the active prompt, then submit plain correction text. Keep
-            # the cancelled request attached so deictic follow-ups ("not that
-            # file") still have an explicit target.
-            interrupted_prompt = ""
-            with state.runtime_lock:
-                if not state.is_running and state.interrupted_prompt_text:
-                    interrupted_prompt = state.interrupted_prompt_text
-                    state.interrupted_prompt_text = ""
-            if interrupted_prompt:
-                user_text = (
-                    f"{interrupted_prompt}\n\n"
-                    f"User correction/guidance after interrupt: {user_text}"
-                )
-                user_content = user_text
 
         # Intercept slash commands — handle locally without calling the LLM.
         # Slash commands are text-only; if the client included images/resources,
@@ -1618,54 +1362,23 @@ class HermesACPAgent(acp.Agent):
                     await self._send_usage_update(state)
                 return PromptResponse(stop_reason="end_turn")
 
-        # If the client sends another regular text prompt while this ACP session
-        # is running, route it through the core active-turn redirect. Rich media
-        # and older runtimes retain the proven next-turn queue fallback.
-        redirected = False
-        queued_depth: int | None = None
+        # If Zed sends another regular prompt while the same ACP session is
+        # still running, queue it instead of racing two AIAgent loops against
+        # the same state.history. /steer and /queue are handled above and can
+        # land immediately.
         with state.runtime_lock:
             if state.is_running:
-                if (
-                    text_only_prompt
-                    and isinstance(user_content, str)
-                    and getattr(
-                        state.agent,
-                        "_supports_active_turn_redirect",
-                        False,
+                queued_text = user_text or "[Image attachment]"
+                state.queued_prompts.append(queued_text)
+                depth = len(state.queued_prompts)
+                if self._conn:
+                    update = acp.update_agent_message_text(
+                        f"Queued for the next turn. ({depth} queued)"
                     )
-                    is True
-                    and hasattr(state.agent, "redirect")
-                ):
-                    try:
-                        redirected = bool(state.agent.redirect(user_content))
-                    except Exception:
-                        logger.debug(
-                            "ACP active-turn redirect failed for %s",
-                            session_id,
-                            exc_info=True,
-                        )
-                if not redirected:
-                    queued_text = user_text or "[Image attachment]"
-                    state.queued_prompts.append(queued_text)
-                    queued_depth = len(state.queued_prompts)
-            else:
-                state.is_running = True
-                state.current_prompt_text = user_text or "[Image attachment]"
-
-        if redirected:
-            if self._conn:
-                update = acp.update_agent_message_text(
-                    "Redirected the active turn with your correction."
-                )
-                await self._conn.session_update(session_id, update)
-            return PromptResponse(stop_reason="end_turn")
-        if queued_depth is not None:
-            if self._conn:
-                update = acp.update_agent_message_text(
-                    f"Queued for the next turn. ({queued_depth} queued)"
-                )
-                await self._conn.session_update(session_id, update)
-            return PromptResponse(stop_reason="end_turn")
+                    await self._conn.session_update(session_id, update)
+                return PromptResponse(stop_reason="end_turn")
+            state.is_running = True
+            state.current_prompt_text = user_text or "[Image attachment]"
 
         logger.info("Prompt on session %s: %s", session_id, user_text[:100])
 
@@ -1733,23 +1446,20 @@ class HermesACPAgent(acp.Agent):
         # Approval callback is per-thread (thread-local, GHSA-qg5c-hvr5-hjgr).
         # Set it INSIDE _run_agent so the TLS write happens in the executor
         # thread — setting it here would write to the event-loop thread's TLS,
-        # not the executor's. Interactive routing uses a contextvar in
-        # tools.approval (set_hermes_interactive_context) rather than
-        # os.environ["HERMES_INTERACTIVE"], so concurrent executor workers can't
-        # race on a process-global flag — one session's restore can't drop
-        # another onto the non-interactive auto-approve path mid-run
-        # (GHSA-96vc-wcxf-jjff). The contextvar write is isolated by the
-        # contextvars.copy_context() wrapper around the executor call below.
+        # not the executor's. Also set HERMES_INTERACTIVE so approval.py
+        # takes the CLI-interactive path (which calls the registered
+        # callback via prompt_dangerous_approval) instead of the
+        # non-interactive auto-approve branch (GHSA-96vc-wcxf-jjff).
         # ACP's conn.request_permission maps cleanly to the interactive
         # callback shape — not the gateway-queue HERMES_EXEC_ASK path,
         # which requires a notify_cb registered in _gateway_notify_cbs.
         previous_approval_cb = None
-        interactive_token = None
+        previous_interactive = None
         edit_approval_token = None
         previous_session_id = None
 
         def _run_agent() -> dict:
-            nonlocal previous_approval_cb, interactive_token, edit_approval_token, previous_session_id
+            nonlocal previous_approval_cb, previous_interactive, edit_approval_token, previous_session_id
             # Bind HERMES_SESSION_KEY for this session so per-session caches
             # (e.g. the interactive sudo password cache in tools.terminal_tool)
             # scope to the ACP session rather than leaking across sessions
@@ -1781,10 +1491,9 @@ class HermesACPAgent(acp.Agent):
                 except Exception:
                     logger.debug("Could not set ACP edit approval requester", exc_info=True)
             # Signal to tools.approval that we have an interactive callback
-            # and the non-interactive auto-approve path must not fire. Uses a
-            # contextvar (not os.environ) so concurrent executor workers don't
-            # race on the flag (GHSA-96vc-wcxf-jjff).
-            interactive_token = set_hermes_interactive_context(True)
+            # and the non-interactive auto-approve path must not fire.
+            previous_interactive = os.environ.get("HERMES_INTERACTIVE")
+            os.environ["HERMES_INTERACTIVE"] = "1"
             # Propagate the originating ACP session id to tools that want to
             # tag side-effects with it (e.g. ``kanban_create`` stamps it on
             # the new task so clients can render a per-session board). Save
@@ -1804,9 +1513,11 @@ class HermesACPAgent(acp.Agent):
                 logger.exception("Agent error in session %s", session_id)
                 return {"final_response": f"Error: {e}", "messages": state.history}
             finally:
-                # Restore the interactive contextvar for this context.
-                if interactive_token is not None:
-                    reset_hermes_interactive_context(interactive_token)
+                # Restore HERMES_INTERACTIVE.
+                if previous_interactive is None:
+                    os.environ.pop("HERMES_INTERACTIVE", None)
+                else:
+                    os.environ["HERMES_INTERACTIVE"] = previous_interactive
                 # Restore HERMES_SESSION_ID symmetrically.
                 if previous_session_id is None:
                     os.environ.pop("HERMES_SESSION_ID", None)
@@ -1900,28 +1611,12 @@ class HermesACPAgent(acp.Agent):
                             self._send_session_info_update(session_id),
                         )
 
-                # Snapshot the runtime identity; the validator lets the
-                # background titler skip its LLM call if the session's model
-                # changed before it fires (#19027).
-                _title_model = getattr(state.agent, "model", None)
-                _title_provider = getattr(state.agent, "provider", None)
                 maybe_auto_title(
                     self.session_manager._get_db(),
                     session_id,
                     user_text,
                     final_response,
                     state.history,
-                    main_runtime={
-                        "model": getattr(state.agent, "model", None),
-                        "provider": getattr(state.agent, "provider", None),
-                        "base_url": getattr(state.agent, "base_url", None),
-                        "api_key": getattr(state.agent, "api_key", None),
-                        "api_mode": getattr(state.agent, "api_mode", None),
-                    },
-                    runtime_validator=lambda: (
-                        getattr(state.agent, "model", None) == _title_model
-                        and getattr(state.agent, "provider", None) == _title_provider
-                    ),
                     title_callback=_notify_title_update,
                 )
             except Exception:
@@ -2039,7 +1734,7 @@ class HermesACPAgent(acp.Agent):
             "tools": self._cmd_tools,
             "context": self._cmd_context,
             "reset": self._cmd_reset,
-            "compress": self._cmd_compress,
+            "compact": self._cmd_compact,
             "steer": self._cmd_steer,
             "queue": self._cmd_queue,
             "version": self._cmd_version,
@@ -2109,8 +1804,8 @@ class HermesACPAgent(acp.Agent):
                 return "No tools available."
             lines = [f"Available tools ({len(tools)}):"]
             for t in tools:
-                name = (t.get("function") or {}).get("name", "?")
-                desc = (t.get("function") or {}).get("description", "")
+                name = t.get("function", {}).get("name", "?")
+                desc = t.get("function", {}).get("description", "")
                 # Truncate long descriptions
                 if len(desc) > 80:
                     desc = desc[:77] + "..."
@@ -2181,7 +1876,7 @@ class HermesACPAgent(acp.Agent):
                     lines.append(
                         f"Compression: due now (threshold ~{threshold_tokens:,}"
                         + (f", {threshold_pct:.0f}%" if threshold_pct else "")
-                        + "). Run /compress."
+                        + "). Run /compact."
                     )
                 else:
                     lines.append(
@@ -2194,39 +1889,24 @@ class HermesACPAgent(acp.Agent):
                 lines.append(f"Compression threshold: ~{threshold_tokens:,} tokens")
 
         if getattr(agent, "compression_enabled", True) is False:
-            lines.append(
-                "Auto-compaction is disabled (compression.enabled: false); "
-                "/compress still compresses manually."
-            )
+            lines.append("Compression is disabled for this agent.")
         else:
-            lines.append("Tip: run /compress to compress manually before the threshold.")
+            lines.append("Tip: run /compact to compress manually before the threshold.")
 
         return "\n".join(lines)
 
     def _cmd_reset(self, args: str, state: SessionState) -> str:
         state.history.clear()
-        reset_failed = False
-        try:
-            reset_session_state = getattr(state.agent, "reset_session_state", None)
-            if callable(reset_session_state):
-                reset_session_state()
-        except Exception:
-            reset_failed = True
-            logger.warning("ACP session state reset failed for %s", state.session_id, exc_info=True)
-        finally:
-            self.session_manager.save_session(state.session_id)
-        if reset_failed:
-            return "Conversation history cleared. Agent session state reset failed; see logs."
+        self.session_manager.save_session(state.session_id)
         return "Conversation history cleared."
 
-    def _cmd_compress(self, args: str, state: SessionState) -> str:
+    def _cmd_compact(self, args: str, state: SessionState) -> str:
         if not state.history:
             return "Nothing to compress — conversation is empty."
         try:
             agent = state.agent
-            # No compression_enabled gate: the flag disables *automatic*
-            # compaction only; manual /compress must keep working (matches
-            # the CLI /compress and gateway handlers).
+            if not getattr(agent, "compression_enabled", True):
+                return "Context compression is disabled for this agent."
             if not hasattr(agent, "_compress_context"):
                 return "Context compression not available for this agent."
 
@@ -2251,7 +1931,6 @@ class HermesACPAgent(acp.Agent):
                     getattr(agent, "_cached_system_prompt", "") or "",
                     approx_tokens=approx_tokens,
                     task_id=state.session_id,
-                    force=True,
                 )
             finally:
                 agent._session_db = original_session_db

acp_adapter/session.py

@@ -26,18 +26,31 @@ from typing import Any, Dict, List, Optional
 logger = logging.getLogger(__name__)
 
 
+def _win_path_to_wsl(path: str) -> str | None:
+    """Convert a Windows drive path to its WSL /mnt/<drive>/... equivalent."""
+    match = re.match(r"^([A-Za-z]):[\\/](.*)$", path)
+    if not match:
+        return None
+    drive = match.group(1).lower()
+    tail = match.group(2).replace("\\", "/")
+    return f"/mnt/{drive}/{tail}"
+
+
 def _translate_acp_cwd(cwd: str) -> str:
     """Translate Windows ACP cwd values when Hermes itself is running in WSL.
 
     Windows ACP clients can launch ``hermes acp`` inside WSL while still sending
-    editor workspaces as Windows drive paths (``E:\\Projects``) or
-    ``\\\\wsl.localhost\\`` UNC paths. Store and execute against the POSIX form so
-    agents, tools, and persisted ACP sessions all agree on the usable workspace.
-    Native Linux/macOS keeps the original cwd unchanged.
+    editor workspaces as Windows drive paths such as ``E:\\Projects``. Store
+    and execute against the WSL mount path so agents, tools, and persisted ACP
+    sessions all agree on the usable workspace. Native Linux/macOS keeps the
+    original cwd unchanged.
     """
-    from hermes_constants import translate_cwd_for_wsl_backend
+    from hermes_constants import is_wsl
 
-    return translate_cwd_for_wsl_backend(str(cwd))
+    if not is_wsl():
+        return cwd
+    translated = _win_path_to_wsl(str(cwd))
+    return translated if translated is not None else cwd
 
 
 def _normalize_cwd_for_compare(cwd: str | None) -> str:
@@ -48,9 +61,7 @@ def _normalize_cwd_for_compare(cwd: str | None) -> str:
 
     # Normalize Windows drive paths into the equivalent WSL mount form so
     # ACP history filters match the same workspace across Windows and WSL.
-    from hermes_constants import windows_path_to_wsl
-
-    translated = windows_path_to_wsl(expanded)
+    translated = _win_path_to_wsl(expanded)
     if translated is not None:
         expanded = translated
     elif re.match(r"^/mnt/[A-Za-z]/", expanded):
@@ -450,47 +461,10 @@ class SessionManager:
                 except Exception:
                     logger.debug("Failed to update ACP session metadata", exc_info=True)
 
-            # When the agent owns persistence to this same SessionDB it has
-            # already flushed the live transcript incrementally during
-            # run_conversation (append_message), and it preserves pre-compaction
-            # turns non-destructively via archive_and_compact() — keeping them on
-            # disk as searchable active=0/compacted=1 rows. Calling
-            # replace_messages() here would then be a redundant double-write that
-            # DELETEs exactly those archived rows (and, after a compression-driven
-            # id rotation where agent.session_id no longer equals
-            # state.session_id, clobbers the ended parent transcript) — silent
-            # data loss for any ACP conversation long enough to compress.
-            #
-            # Only fall back to the destructive atomic replace when the agent is
-            # NOT persisting itself to this DB (e.g. a test agent factory, or a
-            # fresh create/fork whose copied history the agent has not flushed
-            # yet). That path still rolls back on a mid-rewrite failure so the
-            # previously persisted conversation survives (salvaged from #13675).
-            agent = state.agent
-            agent_db = getattr(agent, "_session_db", None)
-            agent_owns_persistence = (
-                agent_db is not None
-                and agent_db is db
-                and bool(getattr(agent, "_session_db_created", False))
-            )
-            if not agent_owns_persistence:
-                # Even when the current agent doesn't "own" persistence, the
-                # session on disk may already carry compaction-archived rows —
-                # e.g. after a model switch or a /restore, both of which mint a
-                # fresh agent with _session_db_created=False (so the check above
-                # is False) yet leave the durable archived transcript in place.
-                # A full-history replace would DELETE those archived rows just
-                # like the owned-agent case. Guard against it: when archived
-                # rows exist, replace ONLY the live (active=1) set and leave the
-                # archived turns untouched; otherwise the destructive replace is
-                # safe (fresh create/fork with no archived history to lose).
-                try:
-                    has_archived = db.has_archived_messages(state.session_id)
-                except Exception:
-                    has_archived = False
-                db.replace_messages(
-                    state.session_id, state.history, active_only=has_archived
-                )
+            # Replace stored messages with current history atomically so a
+            # mid-rewrite failure rolls back and the previously persisted
+            # conversation is preserved (salvaged from #13675).
+            db.replace_messages(state.session_id, state.history)
         except Exception:
             logger.warning("Failed to persist ACP session %s", state.session_id, exc_info=True)
 
@@ -534,15 +508,9 @@ class SessionManager:
 
         model = row.get("model") or None
 
-        # Load conversation history. repair_alternation: this restore feeds
-        # LIVE REPLAY — the loaded list becomes the resumed agent's working
-        # conversation. A durable ``user;user`` violation left in state.db would
-        # otherwise re-fire the pre-request defensive repair on every request
-        # for the rest of the session (see hermes_state.get_messages_as_conversation).
+        # Load conversation history.
         try:
-            history = db.get_messages_as_conversation(
-                session_id, repair_alternation=True
-            )
+            history = db.get_messages_as_conversation(session_id)
         except Exception:
             logger.warning("Failed to load messages for ACP session %s", session_id, exc_info=True)
             history = []

acp_adapter/tools.py

@@ -3,7 +3,6 @@
 from __future__ import annotations
 
 import json
-import logging
 import uuid
 from typing import Any, Dict, List, Optional
 
@@ -15,8 +14,6 @@ from acp.schema import (
     ToolKind,
 )
 
-logger = logging.getLogger(__name__)
-
 # ---------------------------------------------------------------------------
 # Map hermes tool names -> ACP ToolKind
 # ---------------------------------------------------------------------------
@@ -77,7 +74,7 @@ _POLISHED_TOOLS = {
     "kanban_create", "kanban_show", "kanban_comment", "kanban_complete",
     "kanban_block", "kanban_link", "kanban_heartbeat",
     "yb_query_group_info", "yb_query_group_members", "yb_search_sticker",
-    "yb_send_dm", "yb_send_sticker",
+    "yb_send_dm", "yb_send_sticker", "mixture_of_agents",
 }
 
 
@@ -113,12 +110,7 @@ def build_tool_title(tool_name: str, args: Dict[str, Any]) -> str:
     if tool_name == "web_extract":
         urls = args.get("urls", [])
         if urls:
-            first = urls[0]
-            if isinstance(first, dict):
-                first = first.get("url") or first.get("href") or "?"
-            elif not isinstance(first, str):
-                first = "?"
-            return f"extract: {first}" + (f" (+{len(urls)-1})" if len(urls) > 1 else "")
+            return f"extract: {urls[0]}" + (f" (+{len(urls)-1})" if len(urls) > 1 else "")
         return "web extract"
     if tool_name == "process":
         action = str(args.get("action") or "").strip() or "manage"
@@ -387,24 +379,6 @@ def _format_execute_code_result(result: Optional[str]) -> Optional[str]:
     error = str(data.get("error") or "")
     exit_code = data.get("exit_code")
     parts = [f"Exit code: {exit_code}" if exit_code is not None else "Execution complete"]
-    if data.get("stdout_truncated"):
-        total = data.get("stdout_bytes_total")
-        captured = data.get("stdout_bytes_captured")
-        omitted = data.get("stdout_bytes_omitted")
-        if all(isinstance(v, int) for v in (captured, total, omitted)):
-            parts.extend([
-                "",
-                (
-                    "Output truncated: "
-                    f"captured {captured:,} of {total:,} bytes "
-                    f"({omitted:,} omitted)."
-                ),
-            ])
-        else:
-            parts.extend(["", "Output truncated."])
-    warning = str(data.get("warning") or "").strip()
-    if warning:
-        parts.extend(["", "Warning:", warning])
     if output:
         parts.extend(["", "Output:", output])
     if error:
@@ -643,7 +617,7 @@ def _format_session_search_result(result: Optional[str]) -> Optional[str]:
         return None
     mode = data.get("mode") or "search"
     query = data.get("query")
-    lines = ["Recent sessions" if mode == "recent" else "Session search results" + (f" for `{query}`" if query else "")]
+    lines = ["Recent sessions" if mode == "recent" else f"Session search results" + (f" for `{query}`" if query else "")]
     if not results:
         lines.append(str(data.get("message") or "No matching sessions found."))
         return "\n".join(lines)
@@ -1047,37 +1021,7 @@ def build_tool_start(
     *,
     edit_diff: Any = None,
 ) -> ToolCallStart:
-    """Create a ToolCallStart event for the given hermes tool invocation.
-
-    A malformed tool argument (e.g. a non-string ``command``/``path`` from a
-    model that ignores the schema) must never abort the ACP tool-call render —
-    ``build_tool_start`` runs on the live tool-progress callback and during
-    session history replay. On any failure in the title/content/location
-    builders, fall back to a minimal, valid start event. Mirrors
-    ``get_cute_tool_message`` in ``agent/display.py``, wrapped for the same
-    reason on the CLI side.
-    """
-    try:
-        return _build_tool_start(
-            tool_call_id, tool_name, arguments, edit_diff=edit_diff
-        )
-    except Exception as exc:  # noqa: BLE001 — a tool-call render must never abort the turn
-        logger.debug("ACP tool-start render failed for %r: %s", tool_name, exc)
-        safe_name = tool_name if isinstance(tool_name, str) and tool_name else "tool"
-        return acp.start_tool_call(
-            tool_call_id, safe_name, kind=get_tool_kind(safe_name),
-            content=None, locations=[], raw_input=None,
-        )
-
-
-def _build_tool_start(
-    tool_call_id: str,
-    tool_name: str,
-    arguments: Dict[str, Any],
-    *,
-    edit_diff: Any = None,
-) -> ToolCallStart:
-    """Build the ToolCallStart event (unguarded; see ``build_tool_start``)."""
+    """Create a ToolCallStart event for the given hermes tool invocation."""
     kind = get_tool_kind(tool_name)
     title = build_tool_title(tool_name, arguments)
     locations = extract_locations(arguments)

acp_registry/agent.json

@@ -0,0 +1,16 @@
+{
+  "id": "hermes-agent",
+  "name": "Hermes Agent",
+  "version": "0.17.0",
+  "description": "Self-improving open-source AI agent by Nous Research with ACP editor integration, persistent memory, skills, and rich tool support.",
+  "repository": "https://github.com/NousResearch/hermes-agent",
+  "website": "https://hermes-agent.nousresearch.com/docs/user-guide/features/acp",
+  "authors": ["Nous Research"],
+  "license": "MIT",
+  "distribution": {
+    "uvx": {
+      "package": "hermes-agent[acp]==0.17.0",
+      "args": ["hermes-acp"]
+    }
+  }
+}

acp_registry/icon.svg

@@ -0,0 +1,8 @@
+<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 16 16" width="16" height="16" fill="none">
+  <path d="M8 1.5v13" stroke="currentColor" stroke-width="1.5" stroke-linecap="round"/>
+  <path d="M8 3.25c-2.35-1.4-4.7-.95-6.25.35 1.85-.2 3.8.2 5.55 1.55" stroke="currentColor" stroke-width="1.1" stroke-linecap="round" stroke-linejoin="round"/>
+  <path d="M8 3.25c2.35-1.4 4.7-.95 6.25.35-1.85-.2-3.8.2-5.55 1.55" stroke="currentColor" stroke-width="1.1" stroke-linecap="round" stroke-linejoin="round"/>
+  <path d="M8 13.25c-2.3-1-3.05-2.65-1.35-4.15-2 .8-2.35 2.95-.35 4" stroke="currentColor" stroke-width="1.1" stroke-linecap="round" stroke-linejoin="round"/>
+  <path d="M8 13.25c2.3-1 3.05-2.65 1.35-4.15 2 .8 2.35 2.95.35 4" stroke="currentColor" stroke-width="1.1" stroke-linecap="round" stroke-linejoin="round"/>
+  <circle cx="8" cy="1.8" r="1.1" fill="currentColor"/>
+</svg>

agent/account_usage.py

@@ -9,7 +9,7 @@ from typing import TYPE_CHECKING, Any, Optional
 import httpx
 
 from agent.anthropic_adapter import _is_oauth_token, resolve_anthropic_token
-from hermes_cli.auth import AuthError, _read_codex_tokens, resolve_codex_runtime_credentials
+from hermes_cli.auth import _read_codex_tokens, resolve_codex_runtime_credentials
 from hermes_cli.runtime_provider import resolve_runtime_provider
 
 if TYPE_CHECKING:
@@ -214,7 +214,7 @@ def build_nous_credits_snapshot(account_info) -> Optional[AccountUsageSnapshot]:
             return None
 
         details.append(f"Top up: {nous_portal_topup_url(account_info)}")
-        details.append("(or run /topup)")
+        details.append("(or run /credits)")
 
         plan = getattr(sub, "plan", None) if sub is not None else None
         return AccountUsageSnapshot(
@@ -340,7 +340,7 @@ def _snapshot_from_credits_state(state) -> Optional[AccountUsageSnapshot]:
 
 @dataclass(frozen=True)
 class CreditsView:
-    """Surface-agnostic data for the ``/topup`` balance view.
+    """Surface-agnostic data for the ``/credits`` command.
 
     One portal fetch, one parse — consumed identically by the CLI panel, the
     gateway button, and any other money surface. Fail-open: when not logged in
@@ -356,11 +356,11 @@ class CreditsView:
 
 
 def build_credits_view(*, markdown: bool = False, timeout: float = 10.0) -> CreditsView:
-    """Build the /topup balance view: balance block + identity line + top-up URL.
+    """Build the /credits view: balance block + identity line + top-up URL.
 
     Reuses the same account fetch + snapshot + URL builder as the /usage credits
     block, so the numbers always match. The balance block is the rendered
-    snapshot MINUS its trailing top-up/command-hint lines (the /topup surface
+    snapshot MINUS its trailing top-up/command-hint lines (the /credits surface
     supplies its own affordance). Fail-open → ``CreditsView(logged_in=False)``.
     """
     not_logged_in = CreditsView(logged_in=False)
@@ -386,7 +386,7 @@ def build_credits_view(*, markdown: bool = False, timeout: float = 10.0) -> Cred
                 timeout=timeout
             )
     except Exception:
-        logger.debug("credits ▸ /topup portal fetch failed (fail-open)", exc_info=True)
+        logger.debug("credits ▸ /credits portal fetch failed (fail-open)", exc_info=True)
         return not_logged_in
 
     if account is None or not getattr(account, "logged_in", False):
@@ -394,8 +394,8 @@ def build_credits_view(*, markdown: bool = False, timeout: float = 10.0) -> Cred
 
     snapshot = build_nous_credits_snapshot(account)
     # Balance lines = the snapshot block minus the two trailing affordance lines
-    # ("Top up: <url>" + "(or run /topup)") that build_nous_credits_snapshot
-    # appends for the /usage surface. /topup renders its own button/panel.
+    # ("Top up: <url>" + "(or run /credits)") that build_nous_credits_snapshot
+    # appends for the /usage surface. /credits renders its own button/panel.
     balance_lines: list[str] = []
     if snapshot is not None:
         rendered = render_account_usage_lines(snapshot, markdown=markdown)
@@ -425,102 +425,31 @@ def build_credits_view(*, markdown: bool = False, timeout: float = 10.0) -> Cred
     )
 
 
-def _codex_backend_urls(base_url: str) -> tuple[str, str, str]:
-    """Resolve the Codex backend endpoints (usage, reset-credits list, consume).
-
-    Mirrors the Codex CLI's PathStyle split (codex-rs backend-client): base URLs
-    containing ``/backend-api`` use the ChatGPT ``/wham/...`` paths; everything
-    else uses ``/api/codex/...``.
-    """
+def _resolve_codex_usage_url(base_url: str) -> str:
     normalized = (base_url or "").strip().rstrip("/")
     if not normalized:
         normalized = "https://chatgpt.com/backend-api/codex"
     if normalized.endswith("/codex"):
         normalized = normalized[: -len("/codex")]
-    prefix = normalized + ("/wham" if "/backend-api" in normalized else "/api/codex")
-    return (
-        prefix + "/usage",
-        prefix + "/rate-limit-reset-credits",
-        prefix + "/rate-limit-reset-credits/consume",
-    )
+    if "/backend-api" in normalized:
+        return normalized + "/wham/usage"
+    return normalized + "/api/codex/usage"
 
 
-def _resolve_codex_usage_url(base_url: str) -> str:
-    return _codex_backend_urls(base_url)[0]
-
-
-def _resolve_codex_usage_credentials(
-    base_url: Optional[str],
-    api_key: Optional[str],
-) -> tuple[str, str, Optional[str]]:
-    """Resolve Codex quota credentials from the native runtime path.
-
-    Prefer explicit live-agent credentials, then the legacy singleton OAuth
-    state, then the credential pool.  Hermes's native OAuth setup now stores
-    device-code logins in the pool, so quota diagnostics must not depend only
-    on the older singleton store.
-    """
-    explicit_key = str(api_key or "").strip()
-    if explicit_key:
-        return explicit_key, str(base_url or "").strip(), None
-
-    # Tier 2: the native runtime resolver. It ALREADY falls back to the
-    # credential pool when the singleton is empty (see
-    # ``resolve_codex_runtime_credentials`` — issue #32992), so in a pool-only
-    # setup this returns a usable ``source="credential_pool"`` token.
-    #
-    # Only ``AuthError`` ("no creds" / rate-limited) is caught so tier 3 can
-    # run: a broad ``except Exception`` would (a) mask a transient refresh /
-    # network failure and silently hand back a DIFFERENT pool account's usage,
-    # and (b) hide genuine programming errors. A refresh/network error must
-    # propagate — the outer ``fetch_account_usage`` guard fails open (shows
-    # nothing this turn) rather than reporting the wrong account.
-    #
-    # The ``account_id`` (for the ``ChatGPT-Account-Id`` header) is read
-    # best-effort: a partial/missing singleton token store must not sink an
-    # otherwise-usable resolver credential and force a header-less pool fallback.
-    try:
-        creds = resolve_codex_runtime_credentials(refresh_if_expiring=True)
-        account_id: Optional[str] = None
-        try:
-            token_data = _read_codex_tokens()
-            tokens = token_data.get("tokens") or {}
-            account_id = str(tokens.get("account_id", "") or "").strip() or None
-        except AuthError:
-            # Pool-only creds carry no singleton account_id; header is optional.
-            logger.debug("codex ▸ /usage account_id read failed (best-effort)", exc_info=True)
-        return creds["api_key"], str(creds.get("base_url", "") or "").strip(), account_id
-    except AuthError:
-        logger.debug("codex ▸ /usage runtime resolver returned no creds; trying pool", exc_info=True)
-
-    # Tier 3: direct pool select. Reached only when the resolver itself raises
-    # AuthError (e.g. singleton missing AND its own pool read found nothing at
-    # resolve time, but a pool entry is usable now). Pool credentials have no
-    # account_id concept, so the ChatGPT-Account-Id header is intentionally
-    # omitted here.
-    from agent.credential_pool import load_pool
-
-    pool = load_pool("openai-codex")
-    entry = pool.select()
-    if entry is None:
-        raise RuntimeError("No available openai-codex credential in credential pool")
-    return entry.runtime_api_key, str(entry.runtime_base_url or base_url or "").strip(), None
-
-
-def _fetch_codex_account_usage(
-    base_url: Optional[str] = None,
-    api_key: Optional[str] = None,
-) -> Optional[AccountUsageSnapshot]:
-    token, resolved_base_url, account_id = _resolve_codex_usage_credentials(base_url, api_key)
+def _fetch_codex_account_usage() -> Optional[AccountUsageSnapshot]:
+    creds = resolve_codex_runtime_credentials(refresh_if_expiring=True)
+    token_data = _read_codex_tokens()
+    tokens = token_data.get("tokens") or {}
+    account_id = str(tokens.get("account_id", "") or "").strip() or None
     headers = {
-        "Authorization": f"Bearer {token}",
+        "Authorization": f"Bearer {creds['api_key']}",
         "Accept": "application/json",
         "User-Agent": "codex-cli",
     }
     if account_id:
         headers["ChatGPT-Account-Id"] = account_id
     with httpx.Client(timeout=15.0) as client:
-        response = client.get(_resolve_codex_usage_url(resolved_base_url), headers=headers)
+        response = client.get(_resolve_codex_usage_url(creds.get("base_url", "")), headers=headers)
         response.raise_for_status()
     payload = response.json() or {}
     rate_limit = payload.get("rate_limit") or {}
@@ -538,14 +467,6 @@ def _fetch_codex_account_usage(
             )
         )
     details: list[str] = []
-    reset_credits = payload.get("rate_limit_reset_credits") or {}
-    banked = reset_credits.get("available_count")
-    if isinstance(banked, (int, float)) and int(banked) > 0:
-        count = int(banked)
-        plural = "s" if count != 1 else ""
-        details.append(
-            f"You have {count} reset{plural} banked - use /usage reset to activate"
-        )
     credits = payload.get("credits") or {}
     if credits.get("has_credits"):
         balance = credits.get("balance")
@@ -563,191 +484,6 @@ def _fetch_codex_account_usage(
     )
 
 
-@dataclass(frozen=True)
-class CodexResetRedeemResult:
-    """Outcome of a `/usage reset` attempt against the Codex backend."""
-
-    status: str  # reset | nothing_to_reset | no_credit | already_redeemed |
-    #              not_exhausted | no_credits_banked | unavailable
-    message: str
-    available_count: int = 0
-    windows_reset: int = 0
-
-    @property
-    def redeemed(self) -> bool:
-        return self.status == "reset"
-
-
-# Client-side guard threshold: a rate-limit window only counts as exhausted
-# when it is fully used. Below this, redeeming a banked reset wastes most of
-# its value, so we block and point at --force instead.
-_CODEX_WINDOW_EXHAUSTED_PERCENT = 100.0
-
-
-def redeem_codex_reset_credit(
-    *,
-    base_url: Optional[str] = None,
-    api_key: Optional[str] = None,
-    force: bool = False,
-) -> CodexResetRedeemResult:
-    """Redeem one banked Codex rate-limit reset credit (`/usage reset`).
-
-    Flow (mirrors the Codex CLI's reset-credits picker, codex-rs
-    ``backend-client``):
-
-    1. ``GET .../usage`` — read the current windows + banked credit count.
-    2. Guard: zero banked credits → refuse. No window fully used and not
-       ``force`` → refuse with a warning (a banked reset restores the WHOLE
-       5h + weekly allowance; burning it early wastes it). The backend has
-       the same protection (``nothing_to_reset`` doesn't consume the
-       credit), but failing fast client-side gives a clearer message.
-    3. ``POST .../rate-limit-reset-credits/consume`` with a fresh UUID
-       idempotency key (``redeem_request_id``). No ``credit_id`` — the
-       backend picks the next available credit, exactly like the CLI's
-       default "Full reset" option.
-
-    Never raises: every failure mode returns a ``CodexResetRedeemResult``
-    with a user-renderable message.
-    """
-    import uuid
-
-    try:
-        token, resolved_base_url, account_id = _resolve_codex_usage_credentials(base_url, api_key)
-    except Exception:
-        return CodexResetRedeemResult(
-            status="unavailable",
-            message="No Codex credentials available. Run `hermes auth` to sign in with your ChatGPT account.",
-        )
-    usage_url, _credits_url, consume_url = _codex_backend_urls(resolved_base_url)
-    headers = {
-        "Authorization": f"Bearer {token}",
-        "Accept": "application/json",
-        "User-Agent": "codex-cli",
-    }
-    if account_id:
-        headers["ChatGPT-Account-Id"] = account_id
-
-    try:
-        with httpx.Client(timeout=15.0) as client:
-            usage_resp = client.get(usage_url, headers=headers)
-            usage_resp.raise_for_status()
-            payload = usage_resp.json() or {}
-
-            reset_credits = payload.get("rate_limit_reset_credits") or {}
-            raw_count = reset_credits.get("available_count")
-            available = int(raw_count) if isinstance(raw_count, (int, float)) else 0
-            if available <= 0:
-                return CodexResetRedeemResult(
-                    status="no_credits_banked",
-                    message="No banked reset credits on this account — nothing to redeem.",
-                )
-
-            rate_limit = payload.get("rate_limit") or {}
-            worst_used: Optional[float] = None
-            for key in ("primary_window", "secondary_window"):
-                used = (rate_limit.get(key) or {}).get("used_percent")
-                if isinstance(used, (int, float)):
-                    worst_used = max(worst_used or 0.0, float(used))
-            exhausted = worst_used is not None and worst_used >= _CODEX_WINDOW_EXHAUSTED_PERCENT
-            if not exhausted and not force:
-                usage_note = (
-                    f"your busiest window is only {worst_used:.0f}% used"
-                    if worst_used is not None
-                    else "your current usage could not be confirmed as exhausted"
-                )
-                plural = "s" if available != 1 else ""
-                return CodexResetRedeemResult(
-                    status="not_exhausted",
-                    message=(
-                        f"⚠️ Not redeeming: {usage_note}. A banked reset restores your FULL "
-                        f"5h + weekly limits, so spending it now would waste most of it. "
-                        f"You have {available} reset{plural} banked. "
-                        f"Use `/usage reset --force` to redeem anyway."
-                    ),
-                    available_count=available,
-                )
-
-            consume_resp = client.post(
-                consume_url,
-                headers={**headers, "Content-Type": "application/json"},
-                json={"redeem_request_id": str(uuid.uuid4())},
-            )
-            consume_resp.raise_for_status()
-            body = consume_resp.json() or {}
-    except httpx.HTTPStatusError as exc:
-        code = exc.response.status_code
-        if code in (401, 403):
-            return CodexResetRedeemResult(
-                status="unavailable",
-                message=(
-                    "Codex backend rejected the request (HTTP "
-                    f"{code}). Reset credits require ChatGPT-account (OAuth) auth — "
-                    "run `hermes auth` and sign in with your ChatGPT account."
-                ),
-            )
-        return CodexResetRedeemResult(
-            status="unavailable",
-            message=f"Codex backend error (HTTP {code}) — try again shortly.",
-        )
-    except Exception as exc:
-        return CodexResetRedeemResult(
-            status="unavailable",
-            message=f"Could not reach the Codex backend: {exc}",
-        )
-
-    code = str(body.get("code", "") or "").strip().lower()
-    windows_reset = body.get("windows_reset")
-    windows_reset = int(windows_reset) if isinstance(windows_reset, (int, float)) else 0
-    remaining = max(0, available - 1)
-    plural = "s" if remaining != 1 else ""
-    if code == "reset":
-        # The redeemed reset restores the account's quota upstream — lift any
-        # persisted pool cooldowns so Hermes doesn't keep the credential
-        # frozen behind the now-stale ``last_error_reset_at`` (issue #43747).
-        try:
-            from hermes_cli.auth import clear_codex_pool_quota_cooldowns
-
-            clear_codex_pool_quota_cooldowns()
-        except Exception:
-            logger.debug(
-                "Failed to clear Codex pool cooldowns after reset redemption",
-                exc_info=True,
-            )
-        return CodexResetRedeemResult(
-            status="reset",
-            message=(
-                f"✅ Reset redeemed — your usage limits have been reset. "
-                f"{remaining} banked reset{plural} remaining."
-            ),
-            available_count=remaining,
-            windows_reset=windows_reset,
-        )
-    if code == "nothing_to_reset":
-        return CodexResetRedeemResult(
-            status="nothing_to_reset",
-            message=(
-                "Backend reports nothing to reset — your limits aren't exhausted. "
-                "The credit was NOT spent."
-            ),
-            available_count=available,
-        )
-    if code == "no_credit":
-        return CodexResetRedeemResult(
-            status="no_credit",
-            message="Backend reports no available reset credit on this account.",
-        )
-    if code == "already_redeemed":
-        return CodexResetRedeemResult(
-            status="already_redeemed",
-            message="This redemption was already processed — no additional credit was spent.",
-            available_count=remaining,
-        )
-    return CodexResetRedeemResult(
-        status="unavailable",
-        message=f"Unexpected response from the Codex backend: {body!r}",
-    )
-
-
 def _fetch_anthropic_account_usage() -> Optional[AccountUsageSnapshot]:
     token = (resolve_anthropic_token() or "").strip()
     if not token:
@@ -892,7 +628,7 @@ def fetch_account_usage(
         return None
     try:
         if normalized == "openai-codex":
-            return _fetch_codex_account_usage(base_url=base_url, api_key=api_key)
+            return _fetch_codex_account_usage()
         if normalized == "anthropic":
             return _fetch_anthropic_account_usage()
         if normalized == "openrouter":

agent/anthropic_adapter.py

@@ -65,7 +65,6 @@ THINKING_BUDGET = {"xhigh": 32000, "high": 16000, "medium": 8000, "low": 4000}
 # maps to low on every model.  See:
 # https://platform.claude.com/docs/en/about-claude/models/migration-guide
 ADAPTIVE_EFFORT_MAP = {
-    "ultra":   "max",
     "max":     "max",
     "xhigh":   "xhigh",
     "high":    "high",
@@ -127,8 +126,6 @@ _FAST_MODE_SUPPORTED_SUBSTRINGS = ("opus-4-6", "opus-4.6")
 _ANTHROPIC_OUTPUT_LIMITS = {
     # Mythos-class named models (claude-fable-5, …) — 1M context, reasoning
     "claude-fable":      128_000,
-    # Claude Sonnet 5
-    "claude-sonnet-5":   128_000,
     # Claude 4.8
     "claude-opus-4-8":   128_000,
     # Claude 4.7
@@ -249,13 +246,7 @@ def _supports_adaptive_thinking(model: str) -> bool:
     only returns False for the explicit legacy list of older Claude families
     that require manual budget-based thinking. Non-Claude Anthropic-Messages
     models (minimax, qwen3, …) return False so they keep the manual path.
-
-    Kimi / Moonshot models are the exception: their Anthropic-compatible
-    endpoints implement the adaptive contract (``thinking.type="adaptive"``
-    + ``output_config.effort``, including ``xhigh`` and ``display``).
     """
-    if _model_name_is_kimi_family(model):
-        return True
     if not _is_claude_model(model):
         return False
     m = model.lower()
@@ -368,7 +359,7 @@ def _detect_claude_code_version() -> str:
         try:
             result = _sp.run(
                 [cmd, "--version"],
-                capture_output=True, text=True, encoding='utf-8', errors='replace', timeout=5,
+                capture_output=True, text=True, timeout=5,
             )
             if result.returncode == 0 and result.stdout.strip():
                 # Output is like "2.1.74 (Claude Code)" or just "2.1.74"
@@ -457,8 +448,7 @@ def _is_kimi_coding_endpoint(base_url: str | None) -> bool:
 
 # Model-name prefixes that identify the Kimi / Moonshot family.  Covers
 # - official slugs: ``kimi-k2.5``, ``kimi_thinking``, ``moonshot-v1-8k``
-# - common release lines: ``k1.5-...``, ``k2-thinking``, ``k25-...``, ``k2.5-...``,
-#   and the bare Coding Plan slug ``k3`` (plus ``k3.x``/``k3-...`` variants)
+# - common release lines: ``k1.5-...``, ``k2-thinking``, ``k25-...``, ``k2.5-...``
 # Matched case-insensitively against the post-``normalize_model_name`` form,
 # so a caller's ``provider/vendor/model`` slug is handled the same as a
 # bare name.
@@ -468,14 +458,8 @@ _KIMI_FAMILY_MODEL_PREFIXES = (
     "k1.", "k1-",
     "k2.", "k2-",
     "k25", "k2.5",
-    "k3.", "k3-",
 )
 
-# Bare release slugs with no separator suffix (Kimi Coding Plan serves K3
-# as the exact slug ``k3``). Kept exact-match so unrelated model names that
-# merely start with the same characters don't get misclassified.
-_KIMI_FAMILY_EXACT_SLUGS = frozenset({"k3"})
-
 
 def _model_name_is_kimi_family(model: str | None) -> bool:
     if not isinstance(model, str):
@@ -486,8 +470,6 @@ def _model_name_is_kimi_family(model: str | None) -> bool:
     # Strip vendor prefix (e.g. ``moonshotai/kimi-k2.5`` → ``kimi-k2.5``)
     if "/" in m:
         m = m.rsplit("/", 1)[-1]
-    if m in _KIMI_FAMILY_EXACT_SLUGS:
-        return True
     return m.startswith(_KIMI_FAMILY_MODEL_PREFIXES)
 
 
@@ -551,9 +533,8 @@ def _requires_bearer_auth(base_url: str | None) -> bool:
 
     Some third-party /anthropic endpoints implement Anthropic's Messages API but
     require Authorization: Bearer instead of Anthropic's native x-api-key header.
-    MiniMax's global and China Anthropic-compatible endpoints, Azure AI
-    Foundry's Anthropic-style endpoint, and Palantir Foundry's LLM proxy
-    follow this pattern.
+    MiniMax's global and China Anthropic-compatible endpoints, and Azure AI
+    Foundry's Anthropic-style endpoint follow this pattern.
     """
     normalized = _normalize_base_url_text(base_url)
     if not normalized:
@@ -562,11 +543,6 @@ def _requires_bearer_auth(base_url: str | None) -> bool:
     return (
         normalized.startswith(("https://api.minimax.io/anthropic", "https://api.minimaxi.com/anthropic"))
         or "azure.com" in normalized
-        # Palantir Foundry LLM proxy (<org>.palantirfoundry.com/api/v2/llm/proxy/anthropic)
-        # rejects x-api-key with 401 and requires Authorization: Bearer.
-        # Hostname match (not substring) so e.g. evil.com/palantirfoundry
-        # paths don't trigger Bearer auth.
-        or base_url_host_matches(normalized, "palantirfoundry.com")
     )
 
 
@@ -697,9 +673,6 @@ def _build_anthropic_client_with_bearer_hook(
     kwargs = {
         "timeout": timeout_obj,
         "http_client": http_client,
-        # Delegate retry to hermes's outer loop (honors Retry-After); the SDK
-        # default max_retries=2 ignores it and double-retries. (#26293)
-        "max_retries": 0,
         # The SDK requires *something* for api_key/auth_token. Our
         # event hook overrides Authorization per request so this value
         # is never sent. The sentinel string makes accidental leaks
@@ -784,12 +757,6 @@ def build_anthropic_client(
     _read_timeout = timeout if (isinstance(timeout, (int, float)) and timeout > 0) else 900.0
     kwargs = {
         "timeout": Timeout(timeout=float(_read_timeout), connect=10.0),
-        # Delegate all rate-limit / 5xx retry to hermes's outer conversation
-        # loop, which honors Retry-After. The SDK default (max_retries=2) uses
-        # its own 1-2s backoff that ignores Retry-After and double-retries
-        # inside our loop — burning request slots against a bucket that won't
-        # refill for minutes. (#26293)
-        "max_retries": 0,
     }
     if normalized_base_url:
         # Azure Anthropic endpoints require an ``api-version`` query parameter.
@@ -841,7 +808,7 @@ def build_anthropic_client(
         kwargs["auth_token"] = api_key
         kwargs["default_headers"] = {
             "anthropic-beta": ",".join(all_betas),
-            "user-agent": f"claude-code/{_get_claude_code_version()} (external, cli)",
+            "user-agent": f"claude-cli/{_get_claude_code_version()} (external, cli)",
             "x-app": "cli",
         }
     else:
@@ -885,9 +852,6 @@ def build_anthropic_bedrock_client(region: str):
     return _anthropic_sdk.AnthropicBedrock(
         aws_region=region,
         timeout=Timeout(timeout=900.0, connect=10.0),
-        # Delegate retry to hermes's outer loop (honors Retry-After); the SDK
-        # default max_retries=2 ignores it and double-retries. (#26293)
-        max_retries=0,
         default_headers={"anthropic-beta": ",".join([*_COMMON_BETAS, _CONTEXT_1M_BETA])},
     )
 
@@ -914,7 +878,7 @@ def _read_claude_code_credentials_from_keychain() -> Optional[Dict[str, Any]]:
              "-s", "Claude Code-credentials",
              "-w"],
             capture_output=True,
-            text=True, encoding='utf-8', errors='replace',
+            text=True,
             timeout=5,
             stdin=subprocess.DEVNULL,
         )
@@ -950,72 +914,44 @@ def _read_claude_code_credentials_from_keychain() -> Optional[Dict[str, Any]]:
     return None
 
 
-def _read_claude_code_credentials_from_file() -> Optional[Dict[str, Any]]:
-    """Read Claude Code OAuth credentials from ~/.claude/.credentials.json.
-
-    Returns dict with {accessToken, refreshToken?, expiresAt?, source} or None.
-    """
-    cred_path = Path.home() / ".claude" / ".credentials.json"
-    if not cred_path.exists():
-        return None
-    try:
-        data = json.loads(cred_path.read_text(encoding="utf-8"))
-    except (json.JSONDecodeError, OSError, IOError) as e:
-        logger.debug("Failed to read ~/.claude/.credentials.json: %s", e)
-        return None
-
-    oauth_data = data.get("claudeAiOauth")
-    if not (oauth_data and isinstance(oauth_data, dict)):
-        return None
-    access_token = oauth_data.get("accessToken", "")
-    if not access_token:
-        return None
-    return {
-        "accessToken": access_token,
-        "refreshToken": oauth_data.get("refreshToken", ""),
-        "expiresAt": oauth_data.get("expiresAt", 0),
-        "source": "claude_code_credentials_file",
-    }
-
-
 def read_claude_code_credentials() -> Optional[Dict[str, Any]]:
     """Read refreshable Claude Code OAuth credentials.
 
-    Reads from two possible sources and reconciles them:
+    Checks two sources in order:
       1. macOS Keychain (Darwin only) — "Claude Code-credentials" entry
       2. ~/.claude/.credentials.json file
 
-    Selection rules when both are present:
-      - If exactly one is non-expired, prefer that one. (Handles the case
-        where Claude Code refreshes one source but not the other — observed
-        in the wild on Claude Code 2.1.x.)
-      - Otherwise, prefer the source with the later ``expiresAt`` so that
-        any subsequent refresh uses the most recent ``refreshToken``.
-
     This intentionally excludes ~/.claude.json primaryApiKey. Opencode's
     subscription flow is OAuth/setup-token based with refreshable credentials,
     and native direct Anthropic provider usage should follow that path rather
     than auto-detecting Claude's first-party managed key.
 
-    Returns dict with {accessToken, refreshToken?, expiresAt?, source} or None.
+    Returns dict with {accessToken, refreshToken?, expiresAt?} or None.
     """
+    # Try macOS Keychain first (covers Claude Code >=2.1.114)
     kc_creds = _read_claude_code_credentials_from_keychain()
-    file_creds = _read_claude_code_credentials_from_file()
+    if kc_creds:
+        return kc_creds
 
-    if kc_creds and file_creds:
-        kc_valid = is_claude_code_token_valid(kc_creds)
-        file_valid = is_claude_code_token_valid(file_creds)
-        if kc_valid and not file_valid:
-            return kc_creds
-        if file_valid and not kc_valid:
-            return file_creds
-        # Both valid or both expired: prefer the later expiresAt so the
-        # downstream refresh path uses the freshest refresh_token.
-        kc_exp = kc_creds.get("expiresAt", 0) or 0
-        file_exp = file_creds.get("expiresAt", 0) or 0
-        return kc_creds if kc_exp >= file_exp else file_creds
+    # Fall back to JSON file
+    cred_path = Path.home() / ".claude" / ".credentials.json"
+    if cred_path.exists():
+        try:
+            data = json.loads(cred_path.read_text(encoding="utf-8"))
+            oauth_data = data.get("claudeAiOauth")
+            if oauth_data and isinstance(oauth_data, dict):
+                access_token = oauth_data.get("accessToken", "")
+                if access_token:
+                    return {
+                        "accessToken": access_token,
+                        "refreshToken": oauth_data.get("refreshToken", ""),
+                        "expiresAt": oauth_data.get("expiresAt", 0),
+                        "source": "claude_code_credentials_file",
+                    }
+        except (json.JSONDecodeError, OSError, IOError) as e:
+            logger.debug("Failed to read ~/.claude/.credentials.json: %s", e)
 
-    return kc_creds or file_creds
+    return None
 
 
 def is_claude_code_token_valid(creds: Dict[str, Any]) -> bool:
@@ -1069,7 +1005,7 @@ def refresh_anthropic_oauth_pure(refresh_token: str, *, use_json: bool = False)
             data=data,
             headers={
                 "Content-Type": content_type,
-                "User-Agent": _OAUTH_TOKEN_USER_AGENT,
+                "User-Agent": f"claude-cli/{_get_claude_code_version()} (external, cli)",
             },
             method="POST",
         )
@@ -1098,40 +1034,8 @@ def refresh_anthropic_oauth_pure(refresh_token: str, *, use_json: bool = False)
 
 
 def _refresh_oauth_token(creds: Dict[str, Any]) -> Optional[str]:
-    """Attempt to refresh an expired Claude Code OAuth token.
-
-    Claude Code's OAuth refresh tokens are single-use: a successful refresh
-    rotates the pair and invalidates the old refresh token. Claude Code itself
-    also refreshes on its own schedule (IDE/CLI activity), so by the time
-    Hermes notices an expired token, Claude Code may have already rotated it.
-    POSTing our now-stale refresh token in that window races Claude Code and
-    fails with ``invalid_grant``.
-
-    So before refreshing, re-read the live credential sources. If Claude Code
-    has already produced a valid token, adopt it and skip the POST entirely.
-    Only fall back to refreshing ourselves when no fresh credential is found.
-    """
-    # Claude Code may have already refreshed — adopt its token rather than
-    # racing it with our (possibly already-rotated) refresh token. Only adopt
-    # when the live re-read produced a DIFFERENT token with a real future
-    # expiry: re-adopting the same credential we were just handed would be a
-    # no-op, and a 0/absent ``expiresAt`` means "managed key / unknown expiry"
-    # (see is_claude_code_token_valid) which must NOT be treated as a fresh
-    # refresh here.
-    current = read_claude_code_credentials()
-    if current:
-        current_token = current.get("accessToken", "")
-        current_exp = current.get("expiresAt", 0) or 0
-        if (
-            current_token
-            and current_token != creds.get("accessToken", "")
-            and current_exp > 0
-            and is_claude_code_token_valid(current)
-        ):
-            logger.debug("Adopted Claude Code's already-refreshed OAuth token")
-            return current_token
-
-    refresh_token = (current or {}).get("refreshToken", "") or creds.get("refreshToken", "")
+    """Attempt to refresh an expired Claude Code OAuth token."""
+    refresh_token = creds.get("refreshToken", "")
     if not refresh_token:
         logger.debug("No refresh token available — cannot refresh")
         return None
@@ -1255,46 +1159,6 @@ def _prefer_refreshable_claude_code_token(env_token: str, creds: Optional[Dict[s
     return None
 
 
-def _resolve_anthropic_pool_token() -> Optional[str]:
-    """Return the first available Anthropic OAuth token from credential_pool.
-
-    Read-only: enumerates with ``clear_expired=False, refresh=False`` so a bare
-    token *resolve* (which runs from diagnostic/read-only call sites such as
-    ``account_usage`` and ``hermes models``) never mutates ``~/.hermes/auth.json``
-    or makes a network refresh call. Refresh-on-expiry is owned by the API call
-    path's pool recovery, not the resolver.
-    """
-    try:
-        from agent.credential_pool import AUTH_TYPE_OAUTH, load_pool
-    except Exception:
-        return None
-
-    try:
-        pool = load_pool("anthropic")
-        # Enumerate read-only (clear_expired=False, refresh=False): never persist
-        # to auth.json or trigger a network refresh from a bare resolve. select()
-        # is deliberately NOT used — it runs clear_expired=True, refresh=True,
-        # which would violate this read-only contract.
-        entries = pool._available_entries(clear_expired=False, refresh=False)
-    except Exception:
-        logger.debug("Failed to read Anthropic credential_pool", exc_info=True)
-        return None
-
-    for entry in entries:
-        if getattr(entry, "auth_type", None) != AUTH_TYPE_OAUTH:
-            continue
-        # access_token is a declared field but a persisted entry can carry an
-        # explicit null (or a partially-written OAuth entry), so coerce before
-        # strip — a bare None.strip() here would escape the try/excepts above
-        # and crash the whole resolver, taking down the source #5 fallback too.
-        # Matches the aux-client analog (auxiliary_client.py: str(key or "")).
-        token = (getattr(entry, "access_token", None) or "").strip()
-        if token:
-            return token
-
-    return None
-
-
 def resolve_anthropic_token() -> Optional[str]:
     """Resolve an Anthropic token from all available sources.
 
@@ -1303,8 +1167,7 @@ def resolve_anthropic_token() -> Optional[str]:
       2. CLAUDE_CODE_OAUTH_TOKEN env var
       3. Claude Code credentials (~/.claude.json or ~/.claude/.credentials.json)
          — with automatic refresh if expired and a refresh token is available
-      4. Anthropic credential_pool OAuth entry (~/.hermes/auth.json)
-      5. ANTHROPIC_API_KEY env var (regular API key, or legacy fallback)
+      4. ANTHROPIC_API_KEY env var (regular API key, or legacy fallback)
 
     Returns the token string or None.
     """
@@ -1331,12 +1194,7 @@ def resolve_anthropic_token() -> Optional[str]:
     if resolved_claude_token:
         return resolved_claude_token
 
-    # 4. Hermes credential_pool OAuth entry.
-    resolved_pool_token = _resolve_anthropic_pool_token()
-    if resolved_pool_token:
-        return resolved_pool_token
-
-    # 5. Regular API key, or a legacy OAuth token saved in ANTHROPIC_API_KEY.
+    # 4. Regular API key, or a legacy OAuth token saved in ANTHROPIC_API_KEY.
     # This remains as a compatibility fallback for pre-migration Hermes configs.
     api_key = os.getenv("ANTHROPIC_API_KEY", "").strip()
     if api_key:
@@ -1393,29 +1251,10 @@ def run_oauth_setup_token() -> Optional[str]:
 # Stores credentials in ~/.hermes/.anthropic_oauth.json (our own file).
 
 _OAUTH_CLIENT_ID = "9d1c250a-e61b-44d9-88ed-5944d1962f5e"
-# Anthropic migrated the OAuth token endpoint to platform.claude.com;
-# console.anthropic.com now 404s. Callers should iterate _OAUTH_TOKEN_URLS
-# (new host first, console fallback). _OAUTH_TOKEN_URL is kept as the primary
-# for backward compatibility with existing imports and now points at the live host.
-_OAUTH_TOKEN_URLS = [
-    "https://platform.claude.com/v1/oauth/token",
-    "https://console.anthropic.com/v1/oauth/token",
-]
-_OAUTH_TOKEN_URL = _OAUTH_TOKEN_URLS[0]
-# User-Agent sent on the OAuth *token endpoint* (login exchange + refresh).
-# Anthropic rate-limits (HTTP 429) any token-endpoint request whose UA starts
-# with ``claude-code/`` — verified empirically against platform.claude.com:
-# ``claude-code/2.1.200`` and ``Mozilla/5.0`` -> 429; ``axios/*``, ``node``,
-# and SDK-style UAs -> 400 (reached code validation). The real Claude Code CLI
-# exchanges the auth code with a bare axios client (``axios/<ver>``), NOT its
-# ``claude-code/`` inference UA. We mirror that here. NOTE: the *inference* path
-# (build_anthropic_kwargs) still uses the ``claude-code/`` UA + ``x-app: cli`` —
-# that fingerprint is required there and is NOT throttled on the messages API.
-_OAUTH_TOKEN_USER_AGENT = "axios/1.7.9"
+_OAUTH_TOKEN_URL = "https://console.anthropic.com/v1/oauth/token"
 _OAUTH_REDIRECT_URI = "https://console.anthropic.com/oauth/code/callback"
 _OAUTH_SCOPES = "org:create_api_key user:profile user:inference"
-def _get_hermes_oauth_file() -> Path:
-    return get_hermes_home() / ".anthropic_oauth.json"
+_HERMES_OAUTH_FILE = get_hermes_home() / ".anthropic_oauth.json"
 
 
 def _generate_pkce() -> tuple:
@@ -1510,37 +1349,18 @@ def run_hermes_oauth_login_pure() -> Optional[Dict[str, Any]]:
             "code_verifier": verifier,
         }).encode()
 
-        # Anthropic migrated the OAuth token endpoint to platform.claude.com;
-        # console.anthropic.com now 404s. Try the new host first, then fall
-        # back to console for older deployments (mirrors the refresh path).
-        # UA is _OAUTH_TOKEN_USER_AGENT (a non-claude-code UA) — see the
-        # constant's definition for why the token endpoint must not send
-        # claude-code/ (429 UA-prefix block).
-        result = None
-        last_error = None
-        for endpoint in _OAUTH_TOKEN_URLS:
-            req = urllib.request.Request(
-                endpoint,
-                data=exchange_data,
-                headers={
-                    "Content-Type": "application/json",
-                    "User-Agent": _OAUTH_TOKEN_USER_AGENT,
-                },
-                method="POST",
-            )
-            try:
-                with urllib.request.urlopen(req, timeout=15) as resp:
-                    result = json.loads(resp.read().decode())
-                break
-            except Exception as exc:
-                last_error = exc
-                logger.debug("Anthropic token exchange failed at %s: %s", endpoint, exc)
-                continue
+        req = urllib.request.Request(
+            _OAUTH_TOKEN_URL,
+            data=exchange_data,
+            headers={
+                "Content-Type": "application/json",
+                "User-Agent": f"claude-cli/{_get_claude_code_version()} (external, cli)",
+            },
+            method="POST",
+        )
 
-        if result is None:
-            raise last_error if last_error is not None else ValueError(
-                "Anthropic token exchange failed"
-            )
+        with urllib.request.urlopen(req, timeout=15) as resp:
+            result = json.loads(resp.read().decode())
     except Exception as e:
         print(f"Token exchange failed: {e}")
         return None
@@ -1563,10 +1383,9 @@ def run_hermes_oauth_login_pure() -> Optional[Dict[str, Any]]:
 
 def read_hermes_oauth_credentials() -> Optional[Dict[str, Any]]:
     """Read Hermes-managed OAuth credentials from ~/.hermes/.anthropic_oauth.json."""
-    oauth_file = _get_hermes_oauth_file()
-    if oauth_file.exists():
+    if _HERMES_OAUTH_FILE.exists():
         try:
-            data = json.loads(oauth_file.read_text(encoding="utf-8"))
+            data = json.loads(_HERMES_OAUTH_FILE.read_text(encoding="utf-8"))
             if data.get("accessToken"):
                 return data
         except (json.JSONDecodeError, OSError, IOError) as e:
@@ -1591,10 +1410,7 @@ def _is_bedrock_model_id(model: str) -> bool:
     """
     lower = model.lower()
     # Regional inference-profile prefixes
-    if any(lower.startswith(p) for p in (
-        "global.", "us.", "eu.", "apac.", "ap.", "au.", "jp.",
-        "ca.", "sa.", "me.", "af.",
-    )):
+    if any(lower.startswith(p) for p in ("global.", "us.", "eu.", "ap.", "jp.")):
         return True
     # Bare Bedrock model IDs: provider.model-family
     if lower.startswith("anthropic."):
@@ -1881,28 +1697,6 @@ def _content_parts_to_anthropic_blocks(parts: Any) -> List[Dict[str, Any]]:
     return out
 
 
-_EMPTY_TEXT_PLACEHOLDER = "(empty)"
-
-
-def _safe_text(text: Any) -> str:
-    """Return ``text`` if it's non-whitespace, else a non-whitespace placeholder.
-
-    The Anthropic Messages API rejects requests where a text content block is
-    empty or whitespace-only (HTTP 400 "text content blocks must contain
-    non-whitespace text"). When such a block gets stored in session history —
-    e.g. produced by context compression — it is replayed verbatim on every
-    subsequent turn, permanently wedging the session. Coercing to a
-    non-whitespace placeholder is self-healing: the next API call recovers.
-
-    Mirrors ``bedrock_adapter._safe_text`` (#9486); ref #69512.
-    """
-    if text is None:
-        return _EMPTY_TEXT_PLACEHOLDER
-    if not isinstance(text, str):
-        text = str(text)
-    return text if text.strip() else _EMPTY_TEXT_PLACEHOLDER
-
-
 def _sanitize_replay_block(b: Dict[str, Any]) -> Optional[Dict[str, Any]]:
     """Strip output-only fields from a stored Anthropic content block so it is
     valid as REQUEST input on replay.
@@ -1920,18 +1714,7 @@ def _sanitize_replay_block(b: Dict[str, Any]) -> Optional[Dict[str, Any]]:
         return None
     btype = b.get("type")
     if btype == "text":
-        text_val = b.get("text", "")
-        # Bedrock and strict Anthropic-compatible endpoints reject text
-        # blocks where "text" is empty or whitespace-only (#69512). Drop the
-        # blank block (the caller relocates any cache_control it carried and
-        # falls back to a non-whitespace placeholder when nothing survives)
-        # rather than coercing in place — a coerced "(empty)" block would be
-        # model-visible noise next to surviving thinking/tool_use blocks.
-        # Type-safe: captured blocks can carry text=None from an invalid
-        # upstream payload, which a bare .strip() would crash on.
-        if not isinstance(text_val, str) or not text_val.strip():
-            return None
-        out: Dict[str, Any] = {"type": "text", "text": text_val}
+        out: Dict[str, Any] = {"type": "text", "text": b.get("text", "")}
         # citations is input-valid ONLY when it's a non-empty list; the SDK
         # emits citations=None on responses, which the input schema rejects.
         cits = b.get("citations")
@@ -1966,18 +1749,6 @@ def _sanitize_replay_block(b: Dict[str, Any]) -> Optional[Dict[str, Any]]:
     return None
 
 
-def _apply_assistant_cache_control_to_last_cacheable_block(
-    blocks: List[Dict[str, Any]],
-    cache_control: Any,
-) -> None:
-    if not isinstance(cache_control, dict):
-        return
-    for block in reversed(blocks):
-        if isinstance(block, dict) and block.get("type") in {"text", "tool_use"}:
-            block.setdefault("cache_control", dict(cache_control))
-            break
-
-
 def _convert_assistant_message(m: Dict[str, Any]) -> Dict[str, Any]:
     """Convert an assistant message to Anthropic content blocks.
 
@@ -2019,17 +1790,9 @@ def _convert_assistant_message(m: Dict[str, Any]) -> Dict[str, Any]:
                 parsed_args = {}
             redacted_input_by_id[_sanitize_tool_id(tc.get("id", ""))] = parsed_args
         replayed: List[Dict[str, Any]] = []
-        _relocated_replay_cache_control = None
-        _dropped_blank_text = False
         for b in ordered_blocks:
             clean = _sanitize_replay_block(b)
             if clean is None:
-                if isinstance(b, dict) and b.get("type") == "text":
-                    _dropped_blank_text = True
-                if isinstance(b, dict) and isinstance(b.get("cache_control"), dict):
-                    # A dropped blank text block can still carry the cache
-                    # breakpoint marker -- relocate it rather than losing it.
-                    _relocated_replay_cache_control = b["cache_control"]
                 continue
             if clean.get("type") == "tool_use":
                 # Override raw (un-redacted) input with the redacted copy when
@@ -2039,68 +1802,17 @@ def _convert_assistant_message(m: Dict[str, Any]) -> Dict[str, Any]:
                 if redacted is not None:
                     clean["input"] = redacted
             replayed.append(clean)
-        # When every text block was blank and nothing cacheable survived
-        # (e.g. signed thinking + a blank text block, or a SOLE blank
-        # cache-marked block), emit the non-whitespace placeholder so the
-        # replayed message stays schema-valid (#69512) and a relocated cache
-        # marker still has a carrier instead of being silently lost.
-        _has_cacheable_replay = any(
-            isinstance(b, dict) and b.get("type") in {"text", "tool_use"}
-            for b in replayed
-        )
-        if not _has_cacheable_replay and (
-            _dropped_blank_text or _relocated_replay_cache_control is not None
-        ):
-            replayed.append({"type": "text", "text": _EMPTY_TEXT_PLACEHOLDER})
         if replayed:
-            if _relocated_replay_cache_control is not None:
-                _apply_assistant_cache_control_to_last_cacheable_block(
-                    replayed, _relocated_replay_cache_control
-                )
-            _apply_assistant_cache_control_to_last_cacheable_block(
-                replayed, m.get("cache_control")
-            )
             return {"role": "assistant", "content": replayed}
 
     blocks = _extract_preserved_thinking_blocks(m)
-    # Cache markers dropped along with a blank block are relocated onto the
-    # last surviving cacheable block below (via
-    # _apply_assistant_cache_control_to_last_cacheable_block), rather than
-    # lost -- prompt_caching.py's _apply_cache_marker() sets cache_control
-    # directly on content[-1] for list content, so if that last part happens
-    # to be blank text, dropping it silently would lose the breakpoint.
-    _relocated_cache_control = None
     if content:
         if isinstance(content, list):
             converted_content = _convert_content_to_anthropic(content)
             if isinstance(converted_content, list):
-                # Bedrock and strict Anthropic-compatible endpoints reject
-                # text blocks where "text" is empty or whitespace-only. The
-                # ordered-replay path enforces the same invariant via
-                # _sanitize_replay_block(). Type-safe against ANY invalid
-                # "text" value from an upstream payload -- None, or a
-                # truthy non-string like an int -- not just None: checking
-                # isinstance() first (rather than `blk.get("text") or ""`)
-                # means a non-string value is treated as blank/invalid
-                # instead of reaching .strip() and raising AttributeError.
-                for blk in converted_content:
-                    _blk_text = blk.get("text") if isinstance(blk, dict) else None
-                    if (
-                        isinstance(blk, dict)
-                        and blk.get("type") == "text"
-                        and (not isinstance(_blk_text, str) or not _blk_text.strip())
-                    ):
-                        if isinstance(blk.get("cache_control"), dict):
-                            _relocated_cache_control = blk["cache_control"]
-                        continue
-                    blocks.append(blk)
+                blocks.extend(converted_content)
         else:
-            # Scalar (non-list) content: a whitespace-only string is the
-            # same invalid-payload case as an empty list block -- drop it
-            # rather than emitting a blank text block.
-            text_str = str(content)
-            if text_str.strip():
-                blocks.append({"type": "text", "text": text_str})
+            blocks.append({"type": "text", "text": str(content)})
     for tc in m.get("tool_calls", []):
         if not tc or not isinstance(tc, dict):
             continue
@@ -2141,26 +1853,10 @@ def _convert_assistant_message(m: Dict[str, Any]) -> Dict[str, Any]:
     )
     if isinstance(reasoning_content, str) and not _already_has_thinking:
         blocks.insert(0, {"type": "thinking", "thinking": reasoning_content})
-    # Anthropic rejects empty assistant content. IMPORTANT: fall back only
-    # to the placeholder, never to the raw `content` variable -- `content`
-    # is the UNFILTERED original message content, and can itself be exactly
-    # the blank/whitespace-only payload the filtering above just removed
-    # (a sole blank text block, or scalar whitespace with no tool_calls).
-    # `blocks or content` there would silently restore the invalid provider
-    # payload this function exists to prevent (#69512).
-    effective = blocks if blocks else [{"type": "text", "text": _EMPTY_TEXT_PLACEHOLDER}]
-    # Applied here (after the empty-fallback resolution) rather than
-    # earlier against `blocks` directly, so a cache_control relocated from
-    # a dropped blank block that was the ONLY block still lands on the
-    # (empty) placeholder instead of being silently lost when blocks was
-    # empty at the point the marker would otherwise have been applied.
-    if _relocated_cache_control is not None:
-        _apply_assistant_cache_control_to_last_cacheable_block(
-            effective, _relocated_cache_control
-        )
-    _apply_assistant_cache_control_to_last_cacheable_block(
-        effective, m.get("cache_control")
-    )
+    # Anthropic rejects empty assistant content
+    effective = blocks or content
+    if not effective or effective == "":
+        effective = [{"type": "text", "text": "(empty)"}]
     return {"role": "assistant", "content": effective}
 
 
@@ -2231,7 +1927,7 @@ def _convert_user_message(content: Any) -> Dict[str, Any]:
     if isinstance(content, list):
         converted_blocks = _convert_content_to_anthropic(content)
         if not converted_blocks or all(
-            (b.get("text") or "").strip() == ""
+            b.get("text", "").strip() == ""
             for b in converted_blocks
             if isinstance(b, dict) and b.get("type") == "text"
         ):
@@ -2247,81 +1943,57 @@ def _strip_orphaned_tool_blocks(result: List[Dict[str, Any]]) -> None:
     """Strip tool_use blocks with no matching tool_result, and vice versa.
 
     Context compression or session truncation can remove either side of a
-    tool-call pair, or insert messages between a tool_use and its result.
-    Anthropic requires each tool_use to have a matching tool_result in the
-    IMMEDIATELY FOLLOWING user message — a global ID match is not enough.
+    tool-call pair.  Anthropic rejects both orphans with HTTP 400.
+
     Mutates ``result`` in place.
     """
-    # Pass 1: For each assistant message with tool_use blocks, check that
-    # EACH tool_use ID has a matching tool_result in the immediately following
-    # user message.  Strip tool_use blocks that lack an adjacent result —
-    # Anthropic rejects non-adjacent pairs with HTTP 400 even when the IDs
-    # match somewhere later in the conversation.
-    for i, m in enumerate(result):
-        if m.get("role") != "assistant" or not isinstance(m.get("content"), list):
-            continue
-        tool_use_ids_in_turn = {
-            b.get("id")
-            for b in m["content"]
-            if isinstance(b, dict) and b.get("type") == "tool_use"
-        }
-        if not tool_use_ids_in_turn:
-            continue
-
-        # Collect result IDs from the immediately following user message only.
-        adjacent_result_ids: set = set()
-        if i + 1 < len(result):
-            nxt = result[i + 1]
-            if nxt.get("role") == "user" and isinstance(nxt.get("content"), list):
-                for block in nxt["content"]:
-                    if isinstance(block, dict) and block.get("type") == "tool_result":
-                        adjacent_result_ids.add(block.get("tool_use_id"))
-
-        orphaned = tool_use_ids_in_turn - adjacent_result_ids
-        if not orphaned:
-            continue
-
-        kept = [
-            b
-            for b in m["content"]
-            if not (isinstance(b, dict) and b.get("type") == "tool_use" and b.get("id") in orphaned)
-        ]
-        # If stripping an orphaned tool_use mutated a turn that also carries a
-        # signed thinking block, that block's Anthropic signature was computed
-        # against the ORIGINAL (un-stripped) turn content and is now invalid.
-        # Anthropic rejects the replayed turn with HTTP 400 "thinking blocks in
-        # the latest assistant message cannot be modified".  Flag the turn so
-        # _manage_thinking_signatures can demote the dead signature instead of
-        # replaying it verbatim.  See hermes-agent: extended-thinking + parallel
-        # tool batch interrupted mid-flight → non-retryable 400 crash-loop.
-        if len(kept) != len(m["content"]) and any(
-            isinstance(b, dict) and b.get("type") in {"thinking", "redacted_thinking"}
-            for b in m["content"]
-        ):
-            m["_thinking_signature_invalidated"] = True
-        m["content"] = kept if kept else [{"type": "text", "text": "(tool call removed)"}]
-
-    # Pass 2: Rebuild the set of tool_use IDs that survived pass 1, then
-    # strip tool_result blocks that no longer have any matching tool_use
-    # anywhere in the conversation.
-    surviving_tool_use_ids: set = set()
+    # Strip orphaned tool_use blocks (no matching tool_result follows)
+    tool_result_ids = set()
     for m in result:
-        if m.get("role") == "assistant" and isinstance(m.get("content"), list):
+        if m["role"] == "user" and isinstance(m["content"], list):
             for block in m["content"]:
-                if isinstance(block, dict) and block.get("type") == "tool_use":
-                    surviving_tool_use_ids.add(block.get("id"))
-
+                if block.get("type") == "tool_result":
+                    tool_result_ids.add(block.get("tool_use_id"))
     for m in result:
-        if m.get("role") != "user" or not isinstance(m.get("content"), list):
-            continue
-        new_content = [
-            b
-            for b in m["content"]
-            if not (isinstance(b, dict) and b.get("type") == "tool_result")
-            or b.get("tool_use_id") in surviving_tool_use_ids
-        ]
-        if len(new_content) != len(m["content"]):
-            m["content"] = new_content if new_content else [{"type": "text", "text": "(tool result removed)"}]
+        if m["role"] == "assistant" and isinstance(m["content"], list):
+            kept = [
+                b
+                for b in m["content"]
+                if b.get("type") != "tool_use" or b.get("id") in tool_result_ids
+            ]
+            # If stripping an orphaned tool_use mutated a turn that also carries a
+            # signed thinking block, that block's Anthropic signature was computed
+            # against the ORIGINAL (un-stripped) turn content and is now invalid.
+            # Anthropic rejects the replayed turn with HTTP 400 "thinking blocks in
+            # the latest assistant message cannot be modified".  Flag the turn so
+            # _manage_thinking_signatures can demote the dead signature instead of
+            # replaying it verbatim.  See hermes-agent: extended-thinking + parallel
+            # tool batch interrupted mid-flight → non-retryable 400 crash-loop.
+            if len(kept) != len(m["content"]) and any(
+                isinstance(b, dict) and b.get("type") in {"thinking", "redacted_thinking"}
+                for b in m["content"]
+            ):
+                m["_thinking_signature_invalidated"] = True
+            m["content"] = kept
+            if not m["content"]:
+                m["content"] = [{"type": "text", "text": "(tool call removed)"}]
+
+    # Strip orphaned tool_result blocks (no matching tool_use precedes them)
+    tool_use_ids = set()
+    for m in result:
+        if m["role"] == "assistant" and isinstance(m["content"], list):
+            for block in m["content"]:
+                if block.get("type") == "tool_use":
+                    tool_use_ids.add(block.get("id"))
+    for m in result:
+        if m["role"] == "user" and isinstance(m["content"], list):
+            m["content"] = [
+                b
+                for b in m["content"]
+                if b.get("type") != "tool_result" or b.get("tool_use_id") in tool_use_ids
+            ]
+            if not m["content"]:
+                m["content"] = [{"type": "text", "text": "(tool result removed)"}]
 
 
 def _merge_consecutive_roles(result: List[Dict[str, Any]]) -> List[Dict[str, Any]]:
@@ -2398,6 +2070,13 @@ def _manage_thinking_signatures(
     """
     _THINKING_TYPES = frozenset(("thinking", "redacted_thinking"))
     _is_third_party = _is_third_party_anthropic_endpoint(base_url)
+    # Kimi / DeepSeek share a contract: strip signed Anthropic blocks
+    # (neither upstream can validate Anthropic signatures), preserve unsigned
+    # ones synthesised from reasoning_content.  See #13848, #16748.
+    _preserve_unsigned_thinking = (
+        _is_kimi_family_endpoint(base_url, model)
+        or _is_deepseek_anthropic_endpoint(base_url)
+    )
 
     last_assistant_idx = None
     for i in range(len(result) - 1, -1, -1):
@@ -2409,12 +2088,8 @@ def _manage_thinking_signatures(
         if m.get("role") != "assistant" or not isinstance(m.get("content"), list):
             continue
 
-        if _is_kimi_family_endpoint(base_url, model):
-            # Kimi does not enforce thinking signatures — replay as-is
-            # (shared cleanup below still strips cache markers + the internal flag).
-            pass
-        elif _is_deepseek_anthropic_endpoint(base_url):
-            # DeepSeek: strip signed, preserve unsigned.
+        if _preserve_unsigned_thinking:
+            # Kimi / DeepSeek: strip signed, preserve unsigned.
             new_content = []
             for b in m["content"]:
                 if not isinstance(b, dict) or b.get("type") not in _THINKING_TYPES:
@@ -2514,24 +2189,6 @@ def _evict_old_screenshots(result: List[Dict[str, Any]]) -> None:
                 ]
 
 
-def _ensure_leading_user_turn(result: List[Dict[str, Any]]) -> None:
-    """Anthropic requires messages[0] to have role=user.
-
-    After a second context compaction on the auto path the summary can be
-    emitted as role=assistant with nothing in front of it (the system prompt
-    lives outside messages[] or is extracted into the separate ``system``
-    param), so messages[0] ends up assistant and the Messages API rejects
-    the request with HTTP 400 — often masked by a misleading
-    "tool_use ids were found without tool_result blocks" error (#52160).
-
-    Mirror the Bedrock Converse adapter, which unconditionally prepends a
-    minimal user turn when the first message is not user
-    (convert_messages_to_converse).
-    """
-    if result and result[0].get("role") != "user":
-        result.insert(0, {"role": "user", "content": [{"type": "text", "text": " "}]})
-
-
 def convert_messages_to_anthropic(
     messages: List[Dict],
     base_url: str | None = None,
@@ -2590,7 +2247,6 @@ def convert_messages_to_anthropic(
 
     _strip_orphaned_tool_blocks(result)
     result = _merge_consecutive_roles(result)
-    _ensure_leading_user_turn(result)
     _manage_thinking_signatures(result, base_url, model)
     _evict_old_screenshots(result)
 
@@ -2765,19 +2421,25 @@ def build_anthropic_kwargs(
     # MiniMax Anthropic-compat endpoints support thinking (manual mode only,
     # not adaptive).  Haiku does NOT support extended thinking — skip entirely.
     #
-    # Kimi / Moonshot models also use adaptive thinking: their
-    # Anthropic-compatible endpoints (api.moonshot.cn/anthropic,
-    # api.kimi.com/coding) accept ``thinking.type="adaptive"`` +
-    # ``output_config.effort``, and the replay-validation 400s that
-    # originally motivated dropping the parameter (#13848) no longer
-    # occur.  (Kimi on chat_completions enables thinking via extra_body
-    # in the ChatCompletionsTransport — see #13503.)
+    # Kimi's /coding endpoint speaks the Anthropic Messages protocol but has
+    # its own thinking semantics: when ``thinking.enabled`` is sent, Kimi
+    # validates the message history and requires every prior assistant
+    # tool-call message to carry OpenAI-style ``reasoning_content``.  The
+    # Anthropic path never populates that field, and
+    # ``convert_messages_to_anthropic`` strips all Anthropic thinking blocks
+    # on third-party endpoints — so the request fails with HTTP 400
+    # "thinking is enabled but reasoning_content is missing in assistant
+    # tool call message at index N".  Kimi's reasoning is driven server-side
+    # on the /coding route, so skip Anthropic's thinking parameter entirely
+    # for that host.  (Kimi on chat_completions enables thinking via
+    # extra_body in the ChatCompletionsTransport — see #13503.)
     #
     # On 4.7+ the `thinking.display` field defaults to "omitted", which
     # silently hides reasoning text that Hermes surfaces in its CLI. We
     # request "summarized" so the reasoning blocks stay populated — matching
     # 4.6 behavior and preserving the activity-feed UX during long tool runs.
-    if reasoning_config and isinstance(reasoning_config, dict):
+    _is_kimi_coding = _is_kimi_family_endpoint(base_url, model)
+    if reasoning_config and isinstance(reasoning_config, dict) and not _is_kimi_coding:
         if reasoning_config.get("enabled") is not False and "haiku" not in model.lower():
             effort = str(reasoning_config.get("effort", "medium")).lower()
             budget = THINKING_BUDGET.get(effort, 8000)

agent/antigravity_cloudcode_adapter.py

@@ -0,0 +1,164 @@
+"""OpenAI-compatible facade for Antigravity native OAuth inference."""
+
+from __future__ import annotations
+
+from typing import Any, Dict, Iterator, List, Optional
+
+import httpx
+
+from agent import antigravity_oauth
+from agent.antigravity_code_assist import (
+    ANTIGRAVITY_CODE_ASSIST_ENDPOINT,
+    CodeAssistError,
+    ProjectContext,
+    build_headers,
+    resolve_project_context,
+)
+from agent.gemini_cloudcode_adapter import (
+    GeminiCloudCodeClient,
+    _GeminiStreamChunk,
+    _gemini_http_error,
+    _iter_sse_events,
+    _translate_gemini_response,
+    _translate_stream_event,
+    build_gemini_request,
+    wrap_code_assist_request,
+)
+
+MARKER_BASE_URL = "antigravity-pa://google"
+
+
+class AntigravityCloudCodeClient(GeminiCloudCodeClient):
+    """Minimal OpenAI-SDK-compatible facade over Antigravity Code Assist."""
+
+    def __init__(
+        self,
+        *,
+        api_key: Optional[str] = None,
+        base_url: Optional[str] = None,
+        default_headers: Optional[Dict[str, str]] = None,
+        project_id: str = "",
+        **kwargs: Any,
+    ):
+        super().__init__(
+            api_key=api_key or "antigravity-oauth",
+            base_url=base_url or MARKER_BASE_URL,
+            default_headers=default_headers,
+            project_id=project_id,
+            **kwargs,
+        )
+
+    def _ensure_project_context(self, access_token: str, model: str) -> ProjectContext:
+        if self._project_context is not None:
+            return self._project_context  # type: ignore[return-value]
+
+        env_project = antigravity_oauth.resolve_project_id_from_env()
+        creds = antigravity_oauth.load_credentials()
+        stored_project = creds.project_id if creds else ""
+        if stored_project:
+            self._project_context = ProjectContext(
+                project_id=stored_project,
+                managed_project_id=creds.managed_project_id if creds else "",
+                source="stored",
+            )
+            return self._project_context
+
+        ctx = resolve_project_context(
+            access_token,
+            configured_project_id=self._configured_project_id,
+            env_project_id=env_project,
+        )
+        if ctx.project_id or ctx.managed_project_id:
+            antigravity_oauth.update_project_ids(
+                project_id=ctx.project_id,
+                managed_project_id=ctx.managed_project_id,
+            )
+        self._project_context = ctx
+        return ctx
+
+    def _create_chat_completion(
+        self,
+        *,
+        model: str = "gemini-3-flash-agent",
+        messages: Optional[List[Dict[str, Any]]] = None,
+        stream: bool = False,
+        tools: Any = None,
+        tool_choice: Any = None,
+        temperature: Optional[float] = None,
+        max_tokens: Optional[int] = None,
+        top_p: Optional[float] = None,
+        stop: Any = None,
+        extra_body: Optional[Dict[str, Any]] = None,
+        timeout: Any = None,
+        **_: Any,
+    ) -> Any:
+        access_token = antigravity_oauth.get_valid_access_token()
+        ctx = self._ensure_project_context(access_token, model)
+
+        thinking_config = None
+        if isinstance(extra_body, dict):
+            thinking_config = extra_body.get("thinking_config") or extra_body.get("thinkingConfig")
+
+        inner = build_gemini_request(
+            messages=messages or [],
+            tools=tools,
+            tool_choice=tool_choice,
+            temperature=temperature,
+            max_tokens=max_tokens,
+            top_p=top_p,
+            stop=stop,
+            thinking_config=thinking_config,
+        )
+        wrapped = wrap_code_assist_request(
+            project_id=ctx.project_id,
+            model=model,
+            inner_request=inner,
+        )
+
+        headers = build_headers(access_token)
+        headers.update(self._default_headers)
+
+        if stream:
+            return self._stream_completion(model=model, wrapped=wrapped, headers=headers)
+
+        url = f"{ANTIGRAVITY_CODE_ASSIST_ENDPOINT}/v1internal:generateContent"
+        response = self._http.post(url, json=wrapped, headers=headers)
+        if response.status_code != 200:
+            raise _gemini_http_error(response)
+        try:
+            payload = response.json()
+        except ValueError as exc:
+            raise CodeAssistError(
+                f"Invalid JSON from Antigravity Code Assist: {exc}",
+                code="antigravity_code_assist_invalid_json",
+            ) from exc
+        return _translate_gemini_response(payload, model=model)
+
+    def _stream_completion(
+        self,
+        *,
+        model: str,
+        wrapped: Dict[str, Any],
+        headers: Dict[str, str],
+    ) -> Iterator[_GeminiStreamChunk]:
+        url = f"{ANTIGRAVITY_CODE_ASSIST_ENDPOINT}/v1internal:streamGenerateContent?alt=sse"
+        stream_headers = dict(headers)
+        stream_headers["Accept"] = "text/event-stream"
+
+        def _generator() -> Iterator[_GeminiStreamChunk]:
+            try:
+                with self._http.stream("POST", url, json=wrapped, headers=stream_headers) as response:
+                    if response.status_code != 200:
+                        response.read()
+                        raise _gemini_http_error(response)
+                    tool_call_counter: List[int] = [0]
+                    for event in _iter_sse_events(response):
+                        for chunk in _translate_stream_event(event, model, tool_call_counter):
+                            yield chunk
+            except httpx.HTTPError as exc:
+                raise CodeAssistError(
+                    f"Antigravity streaming request failed: {exc}",
+                    code="antigravity_code_assist_stream_error",
+                ) from exc
+
+        return _generator()

agent/antigravity_code_assist.py

@@ -0,0 +1,286 @@
+"""Antigravity Code Assist control-plane helpers.
+
+The new Antigravity CLI uses the same v1internal Code Assist family as
+gemini-cli, but with Antigravity OAuth scopes, metadata and model catalog. This
+module keeps that provider-specific surface separate from
+``agent.google_code_assist``.
+"""
+
+from __future__ import annotations
+
+import json
+import logging
+import urllib.error
+import urllib.request
+import uuid
+from dataclasses import dataclass, field
+from typing import Any, Dict, Iterable, List, Optional
+
+from agent.google_code_assist import CodeAssistError
+
+logger = logging.getLogger(__name__)
+
+ANTIGRAVITY_CODE_ASSIST_ENDPOINT = "https://daily-cloudcode-pa.sandbox.googleapis.com"
+ANTIGRAVITY_MODEL_ENDPOINTS = [
+    ANTIGRAVITY_CODE_ASSIST_ENDPOINT,
+    "https://cloudcode-pa.googleapis.com",
+    "https://autopush-cloudcode-pa.sandbox.googleapis.com",
+]
+
+ANTIGRAVITY_CLIENT_METADATA = {
+    "ideType": "ANTIGRAVITY",
+    "platform": "PLATFORM_UNSPECIFIED",
+    "pluginType": "GEMINI",
+}
+ANTIGRAVITY_USER_AGENT = "antigravity/1.0.0 windows/amd64"
+ANTIGRAVITY_X_GOOG_API_CLIENT = "google-cloud-sdk vscode_cloudshelleditor/0.1"
+
+DEFAULT_AGENT_MODEL_IDS = [
+    "gemini-3-flash-agent",
+    "gemini-3.5-flash-low",
+    "gemini-pro-agent",
+    "gemini-3.1-pro-low",
+    "claude-sonnet-4-6",
+    "claude-opus-4-6-thinking",
+    "gpt-oss-120b-medium",
+]
+
+DEPRECATED_MODEL_REPLACEMENTS = {
+    "gemini-3.1-pro-high": "gemini-pro-agent",
+}
+
+
+@dataclass
+class AntigravityProjectInfo:
+    project_id: str = ""
+    raw: Dict[str, Any] = field(default_factory=dict)
+
+
+@dataclass
+class ProjectContext:
+    project_id: str = ""
+    managed_project_id: str = ""
+    tier_id: str = ""
+    source: str = ""
+
+
+def _client_metadata() -> Dict[str, str]:
+    return dict(ANTIGRAVITY_CLIENT_METADATA)
+
+
+def build_headers(access_token: str, *, accept: str = "application/json") -> Dict[str, str]:
+    return {
+        "Content-Type": "application/json",
+        "Accept": accept,
+        "Authorization": f"Bearer {access_token}",
+        "User-Agent": ANTIGRAVITY_USER_AGENT,
+        "X-Goog-Api-Client": ANTIGRAVITY_X_GOOG_API_CLIENT,
+        "Client-Metadata": json.dumps(_client_metadata(), separators=(",", ":")),
+        "x-activity-request-id": str(uuid.uuid4()),
+    }
+
+
+def _post_json(
+    url: str,
+    body: Dict[str, Any],
+    access_token: str,
+    *,
+    timeout: float = 30.0,
+) -> Dict[str, Any]:
+    data = json.dumps(body).encode("utf-8")
+    request = urllib.request.Request(
+        url,
+        data=data,
+        method="POST",
+        headers=build_headers(access_token),
+    )
+    try:
+        with urllib.request.urlopen(request, timeout=timeout) as response:
+            raw = response.read().decode("utf-8", errors="replace")
+            return json.loads(raw) if raw else {}
+    except urllib.error.HTTPError as exc:
+        detail = ""
+        try:
+            detail = exc.read().decode("utf-8", errors="replace")
+        except Exception:
+            pass
+        raise CodeAssistError(
+            f"Antigravity Code Assist HTTP {exc.code}: {detail or exc.reason}",
+            code=f"antigravity_code_assist_http_{exc.code}",
+        ) from exc
+    except urllib.error.URLError as exc:
+        raise CodeAssistError(
+            f"Antigravity Code Assist request failed: {exc}",
+            code="antigravity_code_assist_network_error",
+        ) from exc
+
+
+def load_code_assist(
+    access_token: str,
+    *,
+    project_id: str = "",
+    endpoint: str = ANTIGRAVITY_CODE_ASSIST_ENDPOINT,
+) -> AntigravityProjectInfo:
+    metadata = _client_metadata()
+    if project_id:
+        metadata["duetProject"] = project_id
+    body: Dict[str, Any] = {"metadata": metadata}
+    if project_id:
+        body["cloudaicompanionProject"] = project_id
+    resp = _post_json(f"{endpoint}/v1internal:loadCodeAssist", body, access_token)
+    project = (
+        str(resp.get("cloudaicompanionProject") or "").strip()
+        or str(resp.get("project") or "").strip()
+    )
+    return AntigravityProjectInfo(project_id=project, raw=resp)
+
+
+def resolve_project_context(
+    access_token: str,
+    *,
+    configured_project_id: str = "",
+    env_project_id: str = "",
+) -> ProjectContext:
+    if configured_project_id:
+        return ProjectContext(project_id=configured_project_id, source="config")
+    if env_project_id:
+        return ProjectContext(project_id=env_project_id, source="env")
+    info = load_code_assist(access_token)
+    if info.project_id:
+        return ProjectContext(
+            project_id=info.project_id,
+            managed_project_id=info.project_id,
+            source="discovered",
+        )
+    # Discovery returned no project (common on fresh consumer accounts that
+    # haven't been onboarded). Fall back to the public default project so the
+    # call chain still succeeds — mirrors the Antigravity CLI reference flow.
+    from agent.antigravity_oauth import DEFAULT_PROJECT_ID
+    return ProjectContext(
+        project_id=DEFAULT_PROJECT_ID,
+        managed_project_id=DEFAULT_PROJECT_ID,
+        source="default",
+    )
+
+
+def fetch_available_models(
+    access_token: str,
+    *,
+    project_id: str = "",
+    endpoint: str = ANTIGRAVITY_CODE_ASSIST_ENDPOINT,
+) -> Dict[str, Any]:
+    body: Dict[str, Any] = {}
+    if project_id:
+        body["project"] = project_id
+    return _post_json(f"{endpoint}/v1internal:fetchAvailableModels", body, access_token)
+
+
+def fetch_available_models_with_fallbacks(
+    access_token: str,
+    *,
+    project_id: str = "",
+    endpoints: Optional[Iterable[str]] = None,
+) -> Dict[str, Any]:
+    last_err: Optional[Exception] = None
+    for endpoint in endpoints or ANTIGRAVITY_MODEL_ENDPOINTS:
+        try:
+            return fetch_available_models(
+                access_token,
+                project_id=project_id,
+                endpoint=endpoint,
+            )
+        except Exception as exc:
+            last_err = exc
+            logger.debug("Antigravity fetchAvailableModels failed on %s: %s", endpoint, exc)
+    if last_err:
+        raise last_err
+    return {}
+
+
+def _model_id_from_value(value: Any) -> str:
+    if isinstance(value, str):
+        return value.strip()
+    if isinstance(value, dict):
+        for key in ("modelId", "model_id", "id", "name"):
+            candidate = str(value.get(key) or "").strip()
+            if candidate:
+                return candidate
+    return ""
+
+
+def _ids_from_sort(sort: Dict[str, Any]) -> List[str]:
+    ids: List[str] = []
+    for key in ("modelIds", "model_ids", "models", "modelSorts"):
+        value = sort.get(key)
+        if isinstance(value, list):
+            for item in value:
+                mid = _model_id_from_value(item)
+                if mid:
+                    ids.append(mid)
+        elif isinstance(value, dict):
+            mid = _model_id_from_value(value)
+            if mid:
+                ids.append(mid)
+    return ids
+
+
+def _is_recommended_sort(sort: Dict[str, Any]) -> bool:
+    label = " ".join(
+        str(sort.get(key) or "")
+        for key in ("name", "displayName", "title", "category", "group")
+    ).lower()
+    return "recommended" in label
+
+
+def _raw_model_ids(payload: Dict[str, Any]) -> List[str]:
+    ids: List[str] = []
+    models = payload.get("models")
+    if isinstance(models, list):
+        for item in models:
+            mid = _model_id_from_value(item)
+            if mid:
+                ids.append(mid)
+    return ids
+
+
+def filter_agent_model_ids(ids: Iterable[str]) -> List[str]:
+    seen: set[str] = set()
+    filtered: List[str] = []
+    raw = [str(mid).strip() for mid in ids if str(mid).strip()]
+    replacements = set(DEPRECATED_MODEL_REPLACEMENTS.values())
+    for mid in raw:
+        if mid in seen:
+            continue
+        if mid.startswith(("chat_", "tab_")):
+            continue
+        if mid in DEPRECATED_MODEL_REPLACEMENTS and DEPRECATED_MODEL_REPLACEMENTS[mid] in raw:
+            continue
+        if mid in replacements and mid in seen:
+            continue
+        seen.add(mid)
+        filtered.append(mid)
+    return filtered
+
+
+def parse_agent_model_ids(payload: Dict[str, Any]) -> List[str]:
+    """Return the user-facing Antigravity agent model list in display order."""
+    sorts = payload.get("agentModelSorts")
+    ordered: List[str] = []
+    if isinstance(sorts, list):
+        recommended = [s for s in sorts if isinstance(s, dict) and _is_recommended_sort(s)]
+        rest = [s for s in sorts if isinstance(s, dict) and not _is_recommended_sort(s)]
+        for sort in recommended + rest:
+            ordered.extend(_ids_from_sort(sort))
+
+    if not ordered:
+        default_id = str(payload.get("defaultAgentModelId") or "").strip()
+        if default_id:
+            ordered.append(default_id)
+        for mid in DEFAULT_AGENT_MODEL_IDS:
+            ordered.append(mid)
+        ordered.extend(_raw_model_ids(payload))
+
+    filtered = filter_agent_model_ids(ordered)
+    if filtered:
+        return filtered
+    return list(DEFAULT_AGENT_MODEL_IDS)

agent/antigravity_oauth.py

@@ -0,0 +1,907 @@
+"""Google OAuth PKCE flow for the Antigravity (google-antigravity) provider.
+
+Tokens are stored separately from the existing ``google-gemini-cli`` provider so
+development and production credentials do not accidentally bleed across:
+
+    ~/.hermes/auth/antigravity_oauth.json
+
+The on-disk schema matches ``agent.google_oauth`` so the runtime resolver can
+share the same refresh/project-id packing convention.
+"""
+
+from __future__ import annotations
+
+import base64
+import contextlib
+import hashlib
+import http.server
+import json
+import logging
+import os
+import re
+import secrets
+import shutil
+import stat
+import threading
+import time
+import urllib.error
+import urllib.parse
+import urllib.request
+import webbrowser
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Any, Dict, Optional, Tuple
+
+from hermes_constants import get_hermes_home
+from utils import atomic_replace
+
+logger = logging.getLogger(__name__)
+
+ENV_CLIENT_ID = "HERMES_ANTIGRAVITY_CLIENT_ID"
+ENV_CLIENT_SECRET = "HERMES_ANTIGRAVITY_CLIENT_SECRET"
+ENV_CLI_PATH = "HERMES_ANTIGRAVITY_CLI_PATH"
+
+# Public Antigravity CLI desktop OAuth client. Like Google's gemini-cli
+# credentials (see agent/google_oauth.py), this is a DESKTOP OAuth client and
+# its "secret" is not confidential — installed-app clients have no
+# secret-keeping requirement (PKCE provides the security), and these creds are
+# baked into every copy of the Antigravity CLI. Shipping them as a fallback
+# lets users without `agy` installed authenticate directly. Split into parts
+# with explicit comments per the convention in google_oauth.py.
+_PUBLIC_CLIENT_ID_PROJECT_NUM = "1071006060591"
+_PUBLIC_CLIENT_ID_HASH = "tmhssin2h21lcre235vtolojh4g403ep"
+_PUBLIC_CLIENT_SECRET_SUFFIX = "K58FWR486LdLJ1mLB8sXC4z6qDAf"
+
+_DEFAULT_CLIENT_ID = (
+    f"{_PUBLIC_CLIENT_ID_PROJECT_NUM}-{_PUBLIC_CLIENT_ID_HASH}"
+    ".apps.googleusercontent.com"
+)
+_DEFAULT_CLIENT_SECRET = f"GOCSPX-{_PUBLIC_CLIENT_SECRET_SUFFIX}"
+
+# Fallback project ID when Code Assist project discovery fails entirely.
+DEFAULT_PROJECT_ID = "rising-fact-p41fc"
+
+_CLIENT_ID_PATTERN = re.compile(
+    r"([0-9]{8,}-[a-z0-9]{20,}\.apps\.googleusercontent\.com)"
+)
+_CLIENT_SECRET_PATTERN = re.compile(r"(GOCSPX-[A-Za-z0-9_-]{20,80})")
+_DISCOVERY_MAX_FILE_BYTES = 25 * 1024 * 1024
+_DISCOVERY_MAX_AGY_BINARY_BYTES = 220 * 1024 * 1024
+_DISCOVERY_MAX_FILES = 600
+_DISCOVERY_EXTENSIONS = {
+    "",
+    ".cjs",
+    ".exe",
+    ".js",
+    ".json",
+    ".mjs",
+    ".node",
+    ".ts",
+}
+_DISCOVERY_SKIP_DIRS = {
+    ".system_generated",
+    "brain",
+    "conversations",
+    "log",
+    "logs",
+    "scratch",
+}
+
+AUTH_ENDPOINT = "https://accounts.google.com/o/oauth2/v2/auth"
+TOKEN_ENDPOINT = "https://oauth2.googleapis.com/token"
+USERINFO_ENDPOINT = "https://www.googleapis.com/oauth2/v1/userinfo"
+
+OAUTH_SCOPES = (
+    "https://www.googleapis.com/auth/cloud-platform "
+    "https://www.googleapis.com/auth/userinfo.email "
+    "https://www.googleapis.com/auth/userinfo.profile "
+    "https://www.googleapis.com/auth/cclog "
+    "https://www.googleapis.com/auth/experimentsandconfigs"
+)
+
+DEFAULT_REDIRECT_PORT = 51121
+REDIRECT_HOST = "localhost"
+CALLBACK_PATH = "/oauth-callback"
+REFRESH_SKEW_SECONDS = 60
+TOKEN_REQUEST_TIMEOUT_SECONDS = 20.0
+CALLBACK_WAIT_SECONDS = 300
+LOCK_TIMEOUT_SECONDS = 30.0
+
+
+class AntigravityOAuthError(RuntimeError):
+    def __init__(self, message: str, *, code: str = "antigravity_oauth_error") -> None:
+        super().__init__(message)
+        self.code = code
+
+
+def _credentials_path() -> Path:
+    return get_hermes_home() / "auth" / "antigravity_oauth.json"
+
+
+def _lock_path() -> Path:
+    return _credentials_path().with_suffix(".json.lock")
+
+
+_lock_state = threading.local()
+
+
+@contextlib.contextmanager
+def _credentials_lock(timeout_seconds: float = LOCK_TIMEOUT_SECONDS):
+    depth = getattr(_lock_state, "depth", 0)
+    if depth > 0:
+        _lock_state.depth = depth + 1
+        try:
+            yield
+        finally:
+            _lock_state.depth -= 1
+        return
+
+    lock_file_path = _lock_path()
+    lock_file_path.parent.mkdir(parents=True, exist_ok=True)
+    fd = os.open(str(lock_file_path), os.O_CREAT | os.O_RDWR, 0o600)
+    acquired = False
+    try:
+        try:
+            import fcntl
+        except ImportError:
+            fcntl = None
+
+        if fcntl is not None:
+            deadline = time.monotonic() + max(0.0, float(timeout_seconds))
+            while True:
+                try:
+                    fcntl.flock(fd, fcntl.LOCK_EX | fcntl.LOCK_NB)
+                    acquired = True
+                    break
+                except BlockingIOError:
+                    if time.monotonic() >= deadline:
+                        raise TimeoutError(
+                            f"Timed out acquiring Antigravity OAuth credentials lock at {lock_file_path}."
+                        )
+                    time.sleep(0.05)
+        else:
+            try:
+                import msvcrt  # type: ignore[import-not-found]
+
+                deadline = time.monotonic() + max(0.0, float(timeout_seconds))
+                while True:
+                    try:
+                        msvcrt.locking(fd, msvcrt.LK_NBLCK, 1)
+                        acquired = True
+                        break
+                    except OSError:
+                        if time.monotonic() >= deadline:
+                            raise TimeoutError(
+                                f"Timed out acquiring Antigravity OAuth credentials lock at {lock_file_path}."
+                            )
+                        time.sleep(0.05)
+            except ImportError:
+                acquired = True
+
+        _lock_state.depth = 1
+        yield
+    finally:
+        try:
+            if acquired:
+                try:
+                    import fcntl
+
+                    fcntl.flock(fd, fcntl.LOCK_UN)
+                except ImportError:
+                    try:
+                        import msvcrt  # type: ignore[import-not-found]
+
+                        try:
+                            msvcrt.locking(fd, msvcrt.LK_UNLCK, 1)
+                        except OSError:
+                            pass
+                    except ImportError:
+                        pass
+        finally:
+            os.close(fd)
+            _lock_state.depth = 0
+
+
+_discovered_creds_cache: Dict[str, Any] = {}
+
+
+def _secret_candidates(raw: str) -> list[str]:
+    candidates: list[str] = []
+    for length in (35, 34, 36, 33, 37, 38, 39, 40, 41, 42):
+        if len(raw) >= length:
+            candidates.append(raw[:length])
+    candidates.append(raw)
+    return list(dict.fromkeys(candidates))
+
+
+def _candidate_discovery_roots() -> list[Path]:
+    roots: list[Path] = []
+
+    explicit = (os.getenv(ENV_CLI_PATH) or "").strip()
+    if explicit:
+        roots.append(Path(explicit))
+
+    for command in ("agy", "agy.exe", "antigravity", "antigravity.exe"):
+        found = shutil.which(command)
+        if found:
+            roots.append(Path(found))
+
+    for env_key in ("LOCALAPPDATA", "APPDATA", "ProgramFiles", "ProgramFiles(x86)"):
+        base = os.getenv(env_key)
+        if not base:
+            continue
+        base_path = Path(base)
+        roots.extend((
+            base_path / "agy",
+            base_path / "agy" / "bin" / "agy.exe",
+            base_path / "Programs" / "Antigravity",
+            base_path / "Programs" / "Antigravity CLI",
+            base_path / "Google" / "Antigravity",
+            base_path / "Google" / "Antigravity CLI",
+        ))
+
+    home = Path.home()
+    for root in (
+        home / ".gemini" / "antigravity-cli",
+        home / ".antigravitycli",
+        home / ".antigravity",
+    ):
+        roots.append(root)
+
+    unique: list[Path] = []
+    seen: set[str] = set()
+    for root in roots:
+        try:
+            key = str(root.expanduser().resolve())
+        except OSError:
+            key = str(root.expanduser())
+        if key not in seen:
+            seen.add(key)
+            unique.append(root)
+    return unique
+
+
+def _iter_discovery_files() -> list[Path]:
+    files: list[Path] = []
+    seen: set[str] = set()
+
+    def add(path: Path) -> None:
+        if len(files) >= _DISCOVERY_MAX_FILES:
+            return
+        if path.suffix.lower() not in _DISCOVERY_EXTENSIONS:
+            return
+        try:
+            stat_info = path.stat()
+            max_bytes = (
+                _DISCOVERY_MAX_AGY_BINARY_BYTES
+                if path.name.lower() in {"agy", "agy.exe", "antigravity", "antigravity.exe"}
+                else _DISCOVERY_MAX_FILE_BYTES
+            )
+            if not path.is_file() or stat_info.st_size > max_bytes:
+                return
+            key = str(path.resolve())
+        except OSError:
+            return
+        if key in seen:
+            return
+        seen.add(key)
+        files.append(path)
+
+    for root in _candidate_discovery_roots():
+        if len(files) >= _DISCOVERY_MAX_FILES:
+            break
+        try:
+            if root.is_file():
+                add(root)
+                continue
+            if not root.is_dir():
+                continue
+        except OSError:
+            continue
+
+        for dirpath, dirnames, filenames in os.walk(root):
+            dirnames[:] = [
+                d for d in dirnames
+                if d not in _DISCOVERY_SKIP_DIRS and not d.startswith(".git")
+            ]
+            for filename in filenames:
+                add(Path(dirpath) / filename)
+                if len(files) >= _DISCOVERY_MAX_FILES:
+                    break
+            if len(files) >= _DISCOVERY_MAX_FILES:
+                break
+    return files
+
+
+def _extract_client_credential_candidates_from_text(content: str) -> list[Tuple[str, str]]:
+    client_ids = list(dict.fromkeys(match.group(1) for match in _CLIENT_ID_PATTERN.finditer(content)))
+    secrets: list[str] = []
+    for match in _CLIENT_SECRET_PATTERN.finditer(content):
+        secrets.extend(_secret_candidates(match.group(1)))
+    secrets = list(dict.fromkeys(secrets))
+    return [(client_id, secret) for client_id in client_ids for secret in secrets]
+
+
+def _discover_client_credentials() -> Tuple[str, str]:
+    if _discovered_creds_cache.get("resolved"):
+        return (
+            _discovered_creds_cache.get("client_id", ""),
+            _discovered_creds_cache.get("client_secret", ""),
+        )
+
+    for path in _iter_discovery_files():
+        try:
+            content = path.read_bytes().decode("utf-8", errors="ignore")
+        except OSError:
+            continue
+        candidates = _extract_client_credential_candidates_from_text(content)
+        if candidates:
+            client_id, client_secret = candidates[0]
+            _discovered_creds_cache.update({
+                "client_id": client_id,
+                "client_secret": client_secret,
+                "candidates": candidates,
+                "resolved": "1",
+            })
+            logger.info("Discovered Antigravity OAuth client credentials from %s", path)
+            return client_id, client_secret
+
+    _discovered_creds_cache["resolved"] = "1"
+    return "", ""
+
+
+def _get_client_id() -> str:
+    env_val = (os.getenv(ENV_CLIENT_ID) or "").strip()
+    if env_val:
+        return env_val
+    discovered, _ = _discover_client_credentials()
+    if discovered:
+        return discovered
+    return _DEFAULT_CLIENT_ID
+
+
+def _get_client_secret() -> str:
+    env_val = (os.getenv(ENV_CLIENT_SECRET) or "").strip()
+    if env_val:
+        return env_val
+    _, discovered = _discover_client_credentials()
+    if discovered:
+        return discovered
+    return _DEFAULT_CLIENT_SECRET
+
+
+def _iter_client_credential_candidates() -> list[Tuple[str, str]]:
+    env_id = (os.getenv(ENV_CLIENT_ID) or "").strip()
+    env_secret = (os.getenv(ENV_CLIENT_SECRET) or "").strip()
+    if env_id and env_secret:
+        return [(env_id, env_secret)]
+
+    _discover_client_credentials()
+    cached = _discovered_creds_cache.get("candidates")
+    candidates: list[Tuple[str, str]] = []
+    if isinstance(cached, list):
+        candidates = [
+            (str(client_id), str(client_secret))
+            for client_id, client_secret in cached
+            if client_id and client_secret
+        ]
+    else:
+        client_id = str(_discovered_creds_cache.get("client_id") or "")
+        client_secret = str(_discovered_creds_cache.get("client_secret") or "")
+        if client_id and client_secret:
+            candidates = [(client_id, client_secret)]
+
+    # Always include the public baked-in default as a last-resort candidate so
+    # users without `agy` installed can still authenticate. De-dupe in case
+    # discovery already surfaced the same client.
+    default_pair = (_DEFAULT_CLIENT_ID, _DEFAULT_CLIENT_SECRET)
+    if default_pair not in candidates:
+        candidates.append(default_pair)
+    return candidates
+
+
+def _require_client_id() -> str:
+    client_id = _get_client_id()
+    if not client_id:
+        raise AntigravityOAuthError(
+            "Antigravity OAuth client ID is not available. Install Antigravity CLI "
+            "so Hermes can discover its desktop OAuth client, set "
+            f"{ENV_CLI_PATH} to the agy executable, or set {ENV_CLIENT_ID} and "
+            f"{ENV_CLIENT_SECRET} in ~/.hermes/.env.",
+            code="antigravity_oauth_client_id_missing",
+        )
+    return client_id
+
+
+def _require_client_secret() -> str:
+    client_secret = _get_client_secret()
+    if not client_secret:
+        raise AntigravityOAuthError(
+            "Antigravity OAuth client secret is not available. Install Antigravity CLI "
+            "so Hermes can discover its desktop OAuth client, set "
+            f"{ENV_CLI_PATH} to the agy executable, or set {ENV_CLIENT_ID} and "
+            f"{ENV_CLIENT_SECRET} in ~/.hermes/.env.",
+            code="antigravity_oauth_client_secret_missing",
+        )
+    return client_secret
+
+
+def _require_client_credentials() -> Tuple[str, str]:
+    candidates = _iter_client_credential_candidates()
+    if not candidates:
+        _require_client_id()
+        _require_client_secret()
+    return candidates[0]
+
+
+def _generate_pkce_pair() -> Tuple[str, str]:
+    verifier = secrets.token_urlsafe(64)
+    digest = hashlib.sha256(verifier.encode("ascii")).digest()
+    challenge = base64.urlsafe_b64encode(digest).rstrip(b"=").decode("ascii")
+    return verifier, challenge
+
+
+@dataclass
+class RefreshParts:
+    refresh_token: str
+    project_id: str = ""
+    managed_project_id: str = ""
+
+    @classmethod
+    def parse(cls, packed: str) -> "RefreshParts":
+        if not packed:
+            return cls(refresh_token="")
+        parts = packed.split("|", 2)
+        return cls(
+            refresh_token=parts[0],
+            project_id=parts[1] if len(parts) > 1 else "",
+            managed_project_id=parts[2] if len(parts) > 2 else "",
+        )
+
+    def format(self) -> str:
+        if not self.refresh_token:
+            return ""
+        if not self.project_id and not self.managed_project_id:
+            return self.refresh_token
+        return f"{self.refresh_token}|{self.project_id}|{self.managed_project_id}"
+
+
+@dataclass
+class AntigravityCredentials:
+    access_token: str
+    refresh_token: str
+    expires_ms: int
+    email: str = ""
+    project_id: str = ""
+    managed_project_id: str = ""
+
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "refresh": RefreshParts(
+                refresh_token=self.refresh_token,
+                project_id=self.project_id,
+                managed_project_id=self.managed_project_id,
+            ).format(),
+            "access": self.access_token,
+            "expires": int(self.expires_ms),
+            "email": self.email,
+        }
+
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "AntigravityCredentials":
+        parts = RefreshParts.parse(str(data.get("refresh", "") or ""))
+        return cls(
+            access_token=str(data.get("access", "") or ""),
+            refresh_token=parts.refresh_token,
+            expires_ms=int(data.get("expires", 0) or 0),
+            email=str(data.get("email", "") or ""),
+            project_id=parts.project_id,
+            managed_project_id=parts.managed_project_id,
+        )
+
+    def access_token_expired(self, skew_seconds: int = REFRESH_SKEW_SECONDS) -> bool:
+        if not self.access_token or not self.expires_ms:
+            return True
+        return (time.time() + max(0, skew_seconds)) * 1000 >= self.expires_ms
+
+
+def load_credentials() -> Optional[AntigravityCredentials]:
+    path = _credentials_path()
+    if not path.exists():
+        return None
+    try:
+        with _credentials_lock():
+            raw = path.read_text(encoding="utf-8")
+        data = json.loads(raw)
+    except (json.JSONDecodeError, OSError, IOError) as exc:
+        logger.warning("Failed to read Antigravity OAuth credentials at %s: %s", path, exc)
+        return None
+    if not isinstance(data, dict):
+        return None
+    creds = AntigravityCredentials.from_dict(data)
+    if not creds.access_token:
+        return None
+    return creds
+
+
+def save_credentials(creds: AntigravityCredentials) -> Path:
+    path = _credentials_path()
+    path.parent.mkdir(parents=True, exist_ok=True)
+    try:
+        os.chmod(path.parent, 0o700)
+    except OSError:
+        pass
+    payload = json.dumps(creds.to_dict(), indent=2, sort_keys=True) + "\n"
+    with _credentials_lock():
+        tmp_path = path.with_suffix(f".tmp.{os.getpid()}.{secrets.token_hex(4)}")
+        try:
+            fd = os.open(
+                str(tmp_path),
+                os.O_WRONLY | os.O_CREAT | os.O_EXCL,
+                stat.S_IRUSR | stat.S_IWUSR,
+            )
+            with os.fdopen(fd, "w", encoding="utf-8") as fh:
+                fh.write(payload)
+                fh.flush()
+                os.fsync(fh.fileno())
+            atomic_replace(tmp_path, path)
+        finally:
+            try:
+                if tmp_path.exists():
+                    tmp_path.unlink()
+            except OSError:
+                pass
+    return path
+
+
+def clear_credentials() -> None:
+    path = _credentials_path()
+    with _credentials_lock():
+        try:
+            path.unlink()
+        except FileNotFoundError:
+            pass
+        except OSError as exc:
+            logger.warning("Failed to remove Antigravity OAuth credentials at %s: %s", path, exc)
+
+
+def _post_form(url: str, data: Dict[str, str], timeout: float) -> Dict[str, Any]:
+    body = urllib.parse.urlencode(data).encode("ascii")
+    request = urllib.request.Request(
+        url,
+        data=body,
+        method="POST",
+        headers={
+            "Content-Type": "application/x-www-form-urlencoded",
+            "Accept": "application/json",
+        },
+    )
+    try:
+        with urllib.request.urlopen(request, timeout=timeout) as response:
+            raw = response.read().decode("utf-8", errors="replace")
+            return json.loads(raw)
+    except urllib.error.HTTPError as exc:
+        detail = ""
+        try:
+            detail = exc.read().decode("utf-8", errors="replace")
+        except Exception:
+            pass
+        code = "antigravity_oauth_token_http_error"
+        if "invalid_grant" in detail.lower():
+            code = "antigravity_oauth_invalid_grant"
+        elif "invalid_client" in detail.lower():
+            code = "antigravity_oauth_invalid_client"
+        raise AntigravityOAuthError(
+            f"Antigravity OAuth token endpoint returned HTTP {exc.code}: {detail or exc.reason}",
+            code=code,
+        ) from exc
+    except urllib.error.URLError as exc:
+        raise AntigravityOAuthError(
+            f"Antigravity OAuth token request failed: {exc}",
+            code="antigravity_oauth_token_network_error",
+        ) from exc
+
+
+def exchange_code(
+    code: str,
+    verifier: str,
+    redirect_uri: str,
+    *,
+    timeout: float = TOKEN_REQUEST_TIMEOUT_SECONDS,
+) -> Dict[str, Any]:
+    last_error: Optional[AntigravityOAuthError] = None
+    candidates = _iter_client_credential_candidates()
+    if not candidates:
+        candidates = [_require_client_credentials()]
+    for client_id, client_secret in candidates:
+        data = {
+            "grant_type": "authorization_code",
+            "code": code,
+            "code_verifier": verifier,
+            "client_id": client_id,
+            "client_secret": client_secret,
+            "redirect_uri": redirect_uri,
+        }
+        try:
+            return _post_form(TOKEN_ENDPOINT, data, timeout)
+        except AntigravityOAuthError as exc:
+            last_error = exc
+            if exc.code != "antigravity_oauth_invalid_client":
+                raise
+    if last_error is not None:
+        raise last_error
+    raise AntigravityOAuthError(
+        "Antigravity OAuth client credentials are unavailable.",
+        code="antigravity_oauth_client_missing",
+    )
+
+
+def refresh_access_token(
+    refresh_token: str,
+    *,
+    timeout: float = TOKEN_REQUEST_TIMEOUT_SECONDS,
+) -> Dict[str, Any]:
+    if not refresh_token:
+        raise AntigravityOAuthError(
+            "Cannot refresh: refresh_token is empty. Re-run OAuth login.",
+            code="antigravity_oauth_refresh_token_missing",
+        )
+    last_error: Optional[AntigravityOAuthError] = None
+    candidates = _iter_client_credential_candidates()
+    if not candidates:
+        candidates = [_require_client_credentials()]
+    for client_id, client_secret in candidates:
+        data = {
+            "grant_type": "refresh_token",
+            "refresh_token": refresh_token,
+            "client_id": client_id,
+            "client_secret": client_secret,
+        }
+        try:
+            return _post_form(TOKEN_ENDPOINT, data, timeout)
+        except AntigravityOAuthError as exc:
+            last_error = exc
+            if exc.code not in {
+                "antigravity_oauth_invalid_client",
+                "antigravity_oauth_invalid_grant",
+            }:
+                raise
+    if last_error is not None:
+        raise last_error
+    raise AntigravityOAuthError(
+        "Antigravity OAuth client credentials are unavailable.",
+        code="antigravity_oauth_client_missing",
+    )
+
+
+def _fetch_user_email(access_token: str, timeout: float = TOKEN_REQUEST_TIMEOUT_SECONDS) -> str:
+    try:
+        request = urllib.request.Request(
+            USERINFO_ENDPOINT + "?alt=json",
+            headers={"Authorization": f"Bearer {access_token}"},
+        )
+        with urllib.request.urlopen(request, timeout=timeout) as response:
+            raw = response.read().decode("utf-8", errors="replace")
+        data = json.loads(raw)
+        return str(data.get("email", "") or "")
+    except Exception as exc:
+        logger.debug("Antigravity userinfo fetch failed (non-fatal): %s", exc)
+        return ""
+
+
+_refresh_inflight: Dict[str, threading.Event] = {}
+_refresh_inflight_lock = threading.Lock()
+
+
+def get_valid_access_token(*, force_refresh: bool = False) -> str:
+    creds = load_credentials()
+    if creds is None:
+        raise AntigravityOAuthError(
+            "No Antigravity OAuth credentials found. Run `hermes login --provider google-antigravity` first.",
+            code="antigravity_oauth_not_logged_in",
+        )
+    if not force_refresh and not creds.access_token_expired():
+        return creds.access_token
+
+    rt = creds.refresh_token
+    with _refresh_inflight_lock:
+        event = _refresh_inflight.get(rt)
+        if event is None:
+            event = threading.Event()
+            _refresh_inflight[rt] = event
+            owner = True
+        else:
+            owner = False
+
+    if not owner:
+        event.wait(timeout=LOCK_TIMEOUT_SECONDS)
+        fresh = load_credentials()
+        if fresh is not None and not fresh.access_token_expired():
+            return fresh.access_token
+
+    try:
+        try:
+            resp = refresh_access_token(rt)
+        except AntigravityOAuthError as exc:
+            if exc.code == "antigravity_oauth_invalid_grant":
+                clear_credentials()
+            raise
+        new_access = str(resp.get("access_token", "") or "").strip()
+        if not new_access:
+            raise AntigravityOAuthError(
+                "Refresh response did not include an access_token.",
+                code="antigravity_oauth_refresh_empty",
+            )
+        creds.access_token = new_access
+        creds.refresh_token = str(resp.get("refresh_token", "") or "").strip() or creds.refresh_token
+        expires_in = int(resp.get("expires_in", 0) or 0)
+        creds.expires_ms = int((time.time() + max(60, expires_in)) * 1000)
+        save_credentials(creds)
+        return creds.access_token
+    finally:
+        if owner:
+            with _refresh_inflight_lock:
+                _refresh_inflight.pop(rt, None)
+            event.set()
+
+
+def update_project_ids(project_id: str = "", managed_project_id: str = "") -> None:
+    creds = load_credentials()
+    if creds is None:
+        return
+    if project_id:
+        creds.project_id = project_id
+    if managed_project_id:
+        creds.managed_project_id = managed_project_id
+    save_credentials(creds)
+
+
+class _OAuthCallbackHandler(http.server.BaseHTTPRequestHandler):
+    expected_state: str = ""
+    captured_code: Optional[str] = None
+    captured_error: Optional[str] = None
+    ready: Optional[threading.Event] = None
+
+    def log_message(self, format: str, *args: Any) -> None:  # noqa: A002, N802
+        logger.debug("Antigravity OAuth callback: " + format, *args)
+
+    def do_GET(self) -> None:  # noqa: N802
+        parsed = urllib.parse.urlparse(self.path)
+        if parsed.path != CALLBACK_PATH:
+            self.send_response(404)
+            self.end_headers()
+            return
+
+        params = urllib.parse.parse_qs(parsed.query)
+        state = (params.get("state") or [""])[0]
+        error = (params.get("error") or [""])[0]
+        code = (params.get("code") or [""])[0]
+
+        handler_cls = type(self)
+        if state != self.expected_state:
+            handler_cls.captured_error = "OAuth state mismatch."
+        elif error:
+            handler_cls.captured_error = error
+        elif not code:
+            handler_cls.captured_error = "OAuth callback did not include a code."
+        else:
+            handler_cls.captured_code = code
+
+        ok = not handler_cls.captured_error
+        self.send_response(200 if ok else 400)
+        self.send_header("Content-Type", "text/html; charset=utf-8")
+        self.end_headers()
+        msg = "Antigravity OAuth complete. You can return to Hermes." if ok else handler_cls.captured_error
+        self.wfile.write(f"<html><body><p>{msg}</p></body></html>".encode("utf-8"))
+        if handler_cls.ready is not None:
+            handler_cls.ready.set()
+
+
+class _ReusableHTTPServer(http.server.HTTPServer):
+    allow_reuse_address = True
+
+
+def resolve_project_id_from_env() -> str:
+    for key in ("HERMES_ANTIGRAVITY_PROJECT_ID", "GOOGLE_CLOUD_PROJECT", "GOOGLE_CLOUD_PROJECT_ID"):
+        value = (os.getenv(key) or "").strip()
+        if value:
+            return value
+    return ""
+
+
+def start_oauth_flow(
+    *,
+    force_relogin: bool = False,
+    open_browser: bool = True,
+    port: int = DEFAULT_REDIRECT_PORT,
+    project_id: str = "",
+) -> AntigravityCredentials:
+    if not force_relogin:
+        existing = load_credentials()
+        if existing and not existing.access_token_expired():
+            return existing
+
+    verifier, challenge = _generate_pkce_pair()
+    state = secrets.token_urlsafe(24)
+    client_id, _ = _require_client_credentials()
+
+    ready = threading.Event()
+    handler_cls = type("AntigravityOAuthCallbackHandler", (_OAuthCallbackHandler,), {})
+    handler_cls.expected_state = state
+    handler_cls.captured_code = None
+    handler_cls.captured_error = None
+    handler_cls.ready = ready
+
+    try:
+        server = _ReusableHTTPServer((REDIRECT_HOST, int(port)), handler_cls)
+    except OSError:
+        server = _ReusableHTTPServer((REDIRECT_HOST, 0), handler_cls)
+    actual_port = int(server.server_address[1])
+    redirect_uri = f"http://{REDIRECT_HOST}:{actual_port}{CALLBACK_PATH}"
+
+    thread = threading.Thread(target=server.serve_forever, daemon=True)
+    thread.start()
+    try:
+        params = {
+            "client_id": client_id,
+            "redirect_uri": redirect_uri,
+            "response_type": "code",
+            "scope": OAUTH_SCOPES,
+            "access_type": "offline",
+            "prompt": "consent",
+            "state": state,
+            "code_challenge": challenge,
+            "code_challenge_method": "S256",
+        }
+        auth_url = AUTH_ENDPOINT + "?" + urllib.parse.urlencode(params)
+        print("Open this URL to authorize Antigravity OAuth:")
+        print(auth_url)
+        if open_browser:
+            webbrowser.open(auth_url)
+        if not ready.wait(timeout=CALLBACK_WAIT_SECONDS):
+            raise AntigravityOAuthError(
+                "Timed out waiting for Antigravity OAuth callback.",
+                code="antigravity_oauth_callback_timeout",
+            )
+        if handler_cls.captured_error:
+            raise AntigravityOAuthError(
+                handler_cls.captured_error,
+                code="antigravity_oauth_callback_error",
+            )
+        code = handler_cls.captured_code or ""
+        token = exchange_code(code, verifier, redirect_uri)
+    finally:
+        server.shutdown()
+        server.server_close()
+
+    access_token = str(token.get("access_token", "") or "").strip()
+    refresh_token = str(token.get("refresh_token", "") or "").strip()
+    if not access_token or not refresh_token:
+        raise AntigravityOAuthError(
+            "Antigravity OAuth response did not include both access_token and refresh_token.",
+            code="antigravity_oauth_missing_token",
+        )
+    expires_in = int(token.get("expires_in", 0) or 0)
+    creds = AntigravityCredentials(
+        access_token=access_token,
+        refresh_token=refresh_token,
+        expires_ms=int((time.time() + max(60, expires_in)) * 1000),
+        email=_fetch_user_email(access_token),
+        project_id=project_id,
+    )
+    save_credentials(creds)
+    return creds
+
+
+def run_antigravity_oauth_login_pure() -> Dict[str, Any]:
+    creds = start_oauth_flow(
+        force_relogin=True,
+        project_id=resolve_project_id_from_env(),
+    )
+    return {
+        "access_token": creds.access_token,
+        "refresh_token": creds.refresh_token,
+        "expires_at_ms": creds.expires_ms,
+        "email": creds.email,
+        "project_id": creds.project_id,
+    }

agent/async_utils.py

@@ -66,19 +66,3 @@ def safe_schedule_threadsafe(
             coro.close()
         log.log(log_level, "%s: %s", log_message, exc)
         return None
-
-
-def consume_detached_task_result(task: "asyncio.Future[Any]") -> None:
-    """Retrieve a detached task's result without surfacing cancellation.
-
-    Used as an ``add_done_callback`` on tasks that were cancelled and
-    detached (e.g. an adapter close path that swallows ``CancelledError``
-    past its teardown deadline). Observing ``task.exception()`` prevents
-    "exception was never retrieved" noise on the event loop; cancellation
-    and any terminal error are deliberately swallowed — the task's owner
-    already gave up on it.
-    """
-    try:
-        task.exception()
-    except (asyncio.CancelledError, Exception):
-        pass

agent/aux_accounting.py

@@ -1,138 +0,0 @@
-"""Ambient session-accounting context for auxiliary LLM calls.
-
-Auxiliary calls (vision, compression, title generation, web_extract,
-session_search, ...) funnel through ``agent.auxiliary_client`` which has no
-session handle — so their token usage was historically discarded, leaving
-dashboard analytics blind to aux model spend (issue #23270).
-
-Instead of threading ``session_db``/``session_id`` parameters through every
-aux call site, the agent loop publishes them here (mirroring the Nous Portal
-conversation context in ``agent.portal_tags``) and the auxiliary client
-records usage at its single response-validation chokepoint.
-
-ContextVar semantics give us the right isolation for free:
-
-* concurrent agents in one process (gateway sessions, delegate subagents)
-  never see each other's accounting context;
-* worker threads spawned via ``tools.thread_context.propagate_context_to_thread``
-  (MoA fan-out, background review) inherit the parent turn's context;
-* asyncio tasks inherit the context of the code that created them.
-
-MoA reference/aggregator slots are explicitly EXCLUDED from recording:
-``agent/conversation_loop.py`` already folds MoA advisor usage and cost into
-the main loop's ``update_token_counts`` delta, so recording them here would
-double-count (see ``_EXCLUDED_TASKS``).
-"""
-
-from __future__ import annotations
-
-import logging
-from contextvars import ContextVar
-from typing import Any, Optional
-
-logger = logging.getLogger(__name__)
-
-# (session_db, session_id) for the active agent turn, or None outside one.
-_accounting: ContextVar[Optional[tuple]] = ContextVar(
-    "aux_accounting_context", default=None
-)
-
-# Aux tasks whose usage is already accounted by the main loop — recording
-# them here would double-count. MoA advisor/aggregator usage is folded into
-# conversation_loop's update_token_counts delta (tokens AND cost).
-_EXCLUDED_TASKS = frozenset({"moa_reference", "moa_aggregator"})
-
-
-def set_accounting_context(session_db: Any, session_id: Optional[str]):
-    """Publish the active session's accounting handles for aux usage recording.
-
-    Called by the agent loop at turn entry. Returns the ContextVar token so
-    callers can ``reset_accounting_context(token)`` on turn exit. Publishing
-    ``None`` handles (no DB / no session id) clears the context.
-    """
-    if session_db is None or not session_id:
-        return _accounting.set(None)
-    return _accounting.set((session_db, session_id))
-
-
-def reset_accounting_context(token) -> None:
-    """Restore the previous accounting context (pair with ``set_...``)."""
-    try:
-        _accounting.reset(token)
-    except Exception:
-        _accounting.set(None)
-
-
-def get_accounting_context() -> Optional[tuple]:
-    """Return ``(session_db, session_id)`` for the active turn, or ``None``."""
-    return _accounting.get()
-
-
-def record_aux_usage(
-    response: Any,
-    task: Optional[str],
-    *,
-    provider: Optional[str] = None,
-    base_url: Optional[str] = None,
-) -> None:
-    """Record an auxiliary response's token usage against the ambient session.
-
-    Called from the auxiliary client's response-validation chokepoint. Strictly
-    best-effort: any failure is swallowed (accounting must never break an aux
-    call). No-ops when:
-
-    * no accounting context is published (call is outside any agent turn),
-    * the task is main-loop-accounted (MoA slots — see ``_EXCLUDED_TASKS``),
-    * the response carries no usage object.
-
-    The model is read from ``response.model`` (accurate even after the aux
-    client's provider-fallback chains); *provider*/*base_url* reflect the
-    originally-resolved route and are best-effort.
-    """
-    try:
-        if not task or task in _EXCLUDED_TASKS:
-            return
-        ctx = _accounting.get()
-        if ctx is None:
-            return
-        session_db, session_id = ctx
-        raw_usage = getattr(response, "usage", None)
-        if raw_usage is None:
-            return
-
-        from agent.usage_pricing import estimate_usage_cost, normalize_usage
-
-        usage = normalize_usage(raw_usage, provider=provider)
-        if not (
-            usage.input_tokens or usage.output_tokens
-            or usage.cache_read_tokens or usage.cache_write_tokens
-            or usage.reasoning_tokens
-        ):
-            return
-
-        model = str(getattr(response, "model", "") or "") or "unknown"
-        estimated_cost = None
-        try:
-            cost = estimate_usage_cost(
-                model, usage, provider=provider, base_url=base_url
-            )
-            if cost.amount_usd is not None:
-                estimated_cost = float(cost.amount_usd)
-        except Exception:
-            logger.debug("Aux usage cost estimation failed", exc_info=True)
-
-        session_db.record_auxiliary_usage(
-            session_id,
-            task,
-            model=model,
-            billing_provider=provider,
-            billing_base_url=base_url,
-            input_tokens=usage.input_tokens,
-            output_tokens=usage.output_tokens,
-            cache_read_tokens=usage.cache_read_tokens,
-            cache_write_tokens=usage.cache_write_tokens,
-            reasoning_tokens=usage.reasoning_tokens,
-            estimated_cost_usd=estimated_cost,
-        )
-    except Exception:
-        logger.debug("Aux usage recording failed (non-fatal)", exc_info=True)

agent/background_review.py

@@ -18,151 +18,15 @@ for invariants and PR review criteria.
 
 from __future__ import annotations
 
+import contextlib
 import json
 import logging
 import os
 from typing import Any, Dict, List, Optional
 
-from agent.thread_scoped_output import thread_scoped_silence
-
 logger = logging.getLogger(__name__)
 
 
-# ---------------------------------------------------------------------------
-# Background-review aux-model selector + routed digest.
-#
-# The review fork runs on the MAIN model by default ("auto"), replaying the
-# full conversation — already warm in the prompt cache, so cheap cache reads.
-# Optimal and unchanged. A user can route the review to a different, cheaper
-# model via auxiliary.background_review.{provider,model}. A different model
-# cannot reuse the parent's cache (different key), so the fork is cold
-# regardless — replaying the full transcript would just cold-write it. So when
-# (and only when) routed to a different model, we replay a compact DIGEST to
-# minimise cold-written tokens. Same model -> full replay; different model ->
-# digest. That's the whole policy.
-# ---------------------------------------------------------------------------
-
-
-def _resolve_review_runtime(agent: Any) -> Dict[str, Any]:
-    """Resolve provider/model/credentials for the review fork.
-
-    Default (auto / unset / same as parent): inherit the parent's live runtime
-    (with codex_app_server -> codex_responses downgrade). ``routed`` is False —
-    the fork uses the main model and the warm cache, exactly as before. When
-    ``auxiliary.background_review.{provider,model}`` names a concrete model
-    different from the parent's, resolve that runtime and set ``routed=True``.
-    """
-    parent_runtime = agent._current_main_runtime()
-    parent_api_mode = parent_runtime.get("api_mode") or None
-    if parent_api_mode == "codex_app_server":
-        parent_api_mode = "codex_responses"
-    parent = {
-        "provider": agent.provider,
-        "model": agent.model,
-        "api_key": parent_runtime.get("api_key") or None,
-        "base_url": parent_runtime.get("base_url") or None,
-        "api_mode": parent_api_mode,
-        "credential_pool": getattr(agent, "_credential_pool", None),
-        "request_overrides": dict(getattr(agent, "request_overrides", {}) or {}),
-        "max_tokens": getattr(agent, "max_tokens", None),
-        "command": getattr(agent, "acp_command", None),
-        "args": list(getattr(agent, "acp_args", []) or []),
-        "routed": False,
-    }
-    try:
-        from hermes_cli.config import load_config
-        cfg = load_config()
-    except Exception:
-        return parent
-    aux = cfg.get("auxiliary", {}) if isinstance(cfg.get("auxiliary"), dict) else {}
-    task = aux.get("background_review", {}) if isinstance(aux.get("background_review"), dict) else {}
-    task_provider = (str(task.get("provider", "")).strip() or None)
-    task_model = (str(task.get("model", "")).strip() or None)
-    task_base_url = (str(task.get("base_url", "")).strip() or None)
-    task_api_key = (str(task.get("api_key", "")).strip() or None)
-    if not (task_provider and task_provider != "auto" and task_model):
-        return parent
-    if task_provider == (agent.provider or "") and task_model == (agent.model or ""):
-        return parent  # same model/provider as parent -> not routed
-    try:
-        from hermes_cli.runtime_provider import resolve_runtime_provider
-        rp = resolve_runtime_provider(
-            requested=task_provider,
-            target_model=task_model,
-            explicit_api_key=task_api_key,
-            explicit_base_url=task_base_url,
-        )
-        return {
-            "provider": rp.get("provider") or task_provider,
-            "model": rp.get("model") or task_model,
-            "api_key": rp.get("api_key"),
-            "base_url": rp.get("base_url"),
-            "api_mode": rp.get("api_mode"),
-            "credential_pool": rp.get("credential_pool"),
-            "request_overrides": dict(rp.get("request_overrides") or {}),
-            "max_tokens": rp.get("max_output_tokens"),
-            "command": rp.get("command"),
-            "args": list(rp.get("args") or []),
-            "routed": True,
-        }
-    except Exception as e:
-        logger.debug("background-review aux routing failed (%s); using main model", e)
-        return parent
-
-
-def _msg_text(m: Dict) -> str:
-    c = m.get("content")
-    if isinstance(c, str):
-        return c.strip()
-    if isinstance(c, list):
-        return " ".join(b.get("text", "") for b in c if isinstance(b, dict)).strip()
-    return ""
-
-
-def _digest_history(messages_snapshot: List[Dict], tail: int = 24) -> List[Dict]:
-    """Compact replay for the routed (different-model) path only.
-
-    Keeps the recent ``tail`` messages verbatim, collapses older turns into one
-    synthetic user-role digest, preserving role alternation. Used ONLY when
-    routed to a different model (cache cold regardless, so fewer cold-written
-    tokens is a pure win). Never on the main-model path (full replay stays warm).
-    """
-    msgs = list(messages_snapshot or [])
-    if len(msgs) <= tail:
-        return msgs
-    keep = msgs[-tail:]
-    while keep and isinstance(keep[0], dict) and keep[0].get("role") == "tool":
-        tail += 1
-        if len(msgs) <= tail:
-            return msgs
-        keep = msgs[-tail:]
-    old = msgs[:-len(keep)]
-    lines: List[str] = []
-    for m in old:
-        if not isinstance(m, dict):
-            continue
-        role = m.get("role")
-        text = _msg_text(m).replace("\n", " ")
-        if role == "user" and text:
-            lines.append(f"USER: {text[:300]}")
-        elif role == "assistant":
-            tcs = m.get("tool_calls") or []
-            if tcs:
-                names = [(tc.get("function") or {}).get("name", "?") for tc in tcs if isinstance(tc, dict)]
-                lines.append(f"ASSISTANT[tools: {', '.join(names)}]")
-            if text:
-                lines.append(f"ASSISTANT: {text[:200]}")
-    digest = {
-        "role": "user",
-        "content": (
-            "[Earlier conversation digest — older turns summarised to bound the "
-            "review's cold-write cost on the routed aux model. Recent turns "
-            "follow verbatim below.]\n" + "\n".join(lines)
-        ),
-    }
-    return [digest] + keep
-
-
 # Review-prompt strings — used by ``spawn_background_review_thread`` to build
 # the user-message that the forked review agent receives.  AIAgent exposes
 # them as class attributes (``_MEMORY_REVIEW_PROMPT`` etc.) for back-compat;
@@ -459,21 +323,10 @@ def summarize_background_review_actions(
             data = json.loads(msg.get("content", "{}"))
         except (json.JSONDecodeError, TypeError):
             continue
-        # ``data`` may not be a dict — some memory/skill tool responses in
-        # older codepaths or wrapper MCP servers return a top-level JSON
-        # list (e.g. ``[{"success": true, ...}]``) or a scalar.  The original
-        # isinstance check below silently skips non-dict payloads, which
-        # is correct, but ``data.get("_change")`` further down can still
-        # hand back a list and break ``change.get("description", "")``.
-        # Defensively normalize everything through a dict-typed alias so
-        # the rest of the function can stay terse without per-call
-        # ``isinstance`` guards (#59437).
         if not isinstance(data, dict) or not data.get("success"):
             continue
         message = data.get("message", "")
-        detail = call_details.get(tcid) or {}
-        if not isinstance(detail, dict):
-            detail = {}
+        detail = call_details.get(tcid, {})
         target = data.get("target", "") or detail.get("target", "")
         is_skill = detail.get("tool") == "skill_manage"
 
@@ -501,30 +354,12 @@ def summarize_background_review_actions(
             content = detail.get("content", "")
             old_text = detail.get("old_text", "")
             skill_name = detail.get("name", "")
-            # ``operations`` may be anything callable put into the JSON
-            # arguments.  Anything non-iterable that isn't a list[str]
-            # of dicts becomes unusable here, so coerce defensively.
-            ops_raw = detail.get("operations")
-            operations: list = (
-                ops_raw if isinstance(ops_raw, list) else []
-            )
+            operations = detail.get("operations") or []
             max_preview = 120
             if is_skill:
-                # ``_change`` is a free-form dict the skill tool leaves in
-                # the response.  Older / wrapper MCP backends return it
-                # as a list, an int, or a JSON-shaped scalar — normalize
-                # to a dict so the .get() calls downstream don't
-                # AttributeError (#59437).
-                change_raw = data.get("_change")
-                change: dict = (
-                    change_raw if isinstance(change_raw, dict) else {}
-                )
-                old_string = (
-                    change.get("old", "") or detail.get("old_string", "")
-                )
-                new_string = (
-                    change.get("new", "") or detail.get("new_string", "")
-                )
+                change = data.get("_change", {})
+                old_string = change.get("old", "") or detail.get("old_string", "")
+                new_string = change.get("new", "") or detail.get("new_string", "")
                 description = change.get("description", "")
                 if action == "patch" and (old_string or new_string):
                     old_preview = old_string[:80].replace("\n", " ") + (
@@ -545,13 +380,7 @@ def summarize_background_review_actions(
                     actions.append(f"📝 {message}" if message else f"Skill {action}")
             elif operations:
                 for op in operations:
-                    # Each element must be a dict-of-fields; some
-                    # legacy codepaths serialize the entry as a bare
-                    # string and the message dict doesn't exist.  Skip
-                    # non-dict items defensively — they have no
-                    # actionable fields anyway (#59437).
-                    if not isinstance(op, dict):
-                        continue
+                    op = op or {}
                     op_act = op.get("action", "")
                     op_content = (op.get("content") or "")
                     op_old = (op.get("old_text") or "")
@@ -648,15 +477,9 @@ def _run_review_in_thread(
     review_agent = None
     review_messages: List[Dict] = []
     try:
-        # Silence stdout/stderr for THIS worker thread only.  A process-global
-        # ``contextlib.redirect_stdout(devnull)`` here would also blank
-        # ``sys.stdout``/``sys.stderr`` for every other thread — including a
-        # gateway event-loop thread driving a Telegram long-poll — for the full
-        # duration of the review (tens of seconds), swallowing their console
-        # output (#55769 / #55925).  ``thread_scoped_silence`` routes only this
-        # thread's writes to devnull and leaves all other threads on the real
-        # streams.
-        with thread_scoped_silence():
+        with open(os.devnull, "w", encoding="utf-8") as _devnull, \
+             contextlib.redirect_stdout(_devnull), \
+             contextlib.redirect_stderr(_devnull):
             # Inherit the parent agent's live runtime (provider, model,
             # base_url, api_key, api_mode) so the fork uses the exact
             # same credentials the main turn is using.  Without this,
@@ -665,13 +488,18 @@ def _run_review_in_thread(
             # creds, or credential-pool setups where the resolver can't
             # reconstruct auth from scratch -- producing the spurious
             # "No LLM provider configured" warning at end of turn.
-            # _resolve_review_runtime() returns the parent's live runtime by
-            # default (routed=False; main model, warm cache), or — when the user
-            # set auxiliary.background_review.{provider,model} to a different
-            # model — that model's runtime (routed=True). The codex_app_server
-            # -> codex_responses downgrade is applied inside the resolver.
-            _rt = _resolve_review_runtime(agent)
-            _routed = bool(_rt.get("routed"))
+            _parent_runtime = agent._current_main_runtime()
+            _parent_api_mode = _parent_runtime.get("api_mode") or None
+            # The review fork needs to call agent-loop tools (memory,
+            # skill_manage). Those tools require Hermes' own dispatch,
+            # which the codex_app_server runtime bypasses entirely
+            # (it runs the turn inside codex's subprocess). So when
+            # the parent is on codex_app_server, downgrade the review
+            # fork to codex_responses — same auth/credentials, but
+            # talks to the OpenAI Responses API directly so Hermes
+            # owns the loop and the agent-loop tools dispatch.
+            if _parent_api_mode == "codex_app_server":
+                _parent_api_mode = "codex_responses"
             # skip_memory=True keeps the review fork from
             # touching external memory plugins (honcho, mem0,
             # supermemory, etc.).  Without it, the fork's
@@ -690,41 +518,20 @@ def _run_review_in_thread(
             # Match parent's toolset config so ``tools[]`` is byte-identical
             # in the request body — Anthropic's cache key includes it.
             # (The runtime whitelist below still restricts dispatch.)
-            _fork_kwargs: Dict[str, Any] = {}
-            if isinstance(_rt.get("max_tokens"), int):
-                _fork_kwargs["max_tokens"] = _rt["max_tokens"]
-            if isinstance(_rt.get("command"), str) and _rt["command"]:
-                _fork_kwargs["acp_command"] = _rt["command"]
-                _fork_kwargs["acp_args"] = _rt.get("args") or []
-            # Match parent's reasoning config so the fork's ``thinking`` /
-            # ``output_config`` are byte-identical in the request body —
-            # Anthropic's cache key is namespaced by ``thinking`` presence.
-            # Same-model path only: when routed to a different aux model the
-            # cache is cold regardless (parity buys nothing) and the parent's
-            # effort vocabulary may not be valid for the routed model/provider
-            # (e.g. OpenRouter ``extra_body.reasoning.effort`` is forwarded
-            # unclamped; codex_responses passes ``max``/``ultra`` through
-            # unmapped except on gpt-5.6/xAI). Let the routed fork use
-            # provider defaults — matching the ``not _routed`` gate on
-            # _cached_system_prompt below.
-            if not _routed:
-                _fork_kwargs["reasoning_config"] = getattr(agent, "reasoning_config", None)
             review_agent = AIAgent(
-                model=_rt.get("model") or agent.model,
+                model=agent.model,
                 max_iterations=16,
                 quiet_mode=True,
                 platform=agent.platform,
-                provider=_rt.get("provider") or agent.provider,
-                api_mode=_rt.get("api_mode"),
-                base_url=_rt.get("base_url") or None,
-                api_key=_rt.get("api_key") or None,
-                credential_pool=_rt.get("credential_pool"),
-                request_overrides=_rt.get("request_overrides") or {},
+                provider=agent.provider,
+                api_mode=_parent_api_mode,
+                base_url=_parent_runtime.get("base_url") or None,
+                api_key=_parent_runtime.get("api_key") or None,
+                credential_pool=getattr(agent, "_credential_pool", None),
                 parent_session_id=agent.session_id,
                 enabled_toolsets=getattr(agent, "enabled_toolsets", None),
                 disabled_toolsets=getattr(agent, "disabled_toolsets", None),
                 skip_memory=True,
-                **_fork_kwargs,
             )
             review_agent._memory_write_origin = "background_review"
             review_agent._memory_write_context = "background_review"
@@ -740,20 +547,6 @@ def _run_review_in_thread(
             review_agent._user_profile_enabled = agent._user_profile_enabled
             review_agent._memory_nudge_interval = 0
             review_agent._skill_nudge_interval = 0
-            # PERSISTENCE ISOLATION (the curator-takeover root cause): the fork
-            # shares the parent's session_id (set below, for prompt-cache
-            # warmth), so without this it would write its harness turn ("Review
-            # the conversation above and update the skill library…") + its own
-            # response straight into the user's REAL session in state.db. On the
-            # user's next live turn the agent re-reads that injected user message
-            # as a standing instruction and "becomes" the curator, refusing the
-            # actual task. _persist_disabled hard-stops every DB write/lazy-open
-            # path (_flush_messages_to_session_db, _ensure_db_session,
-            # _get_session_db_for_recall); the review writes only to the skill
-            # and memory stores via its tools, which is all it needs.
-            review_agent._persist_disabled = True
-            review_agent._session_db = None
-            review_agent._session_json_enabled = False
             # Suppress all status/warning emits from the fork so the
             # user only sees the final successful-action summary.
             # Without this, mid-review "Iteration budget exhausted",
@@ -772,20 +565,15 @@ def _run_review_in_thread(
             # issue #25322 and PR #17276 for the full analysis +
             # measured impact (~26% end-to-end cost reduction on
             # Sonnet 4.5).
-            # Share the parent's warm cached system prompt ONLY when the review
-            # runs on the SAME model (not routed). When routed to a different
-            # model the parent's cached prompt is for the wrong model/cache key
-            # and would miss anyway, so let the routed fork build its own.
-            if not _routed:
-                review_agent._cached_system_prompt = agent._cached_system_prompt
-                # Defensive: pin session_start + session_id to the
-                # parent's so any code path that re-renders parts of
-                # the system prompt (compression, plugin hooks) still
-                # produces byte-identical output. The cached-prompt
-                # assignment above already short-circuits the normal
-                # rebuild path, but these pins guarantee parity even
-                # if a future code path bypasses the cache.
-                review_agent.session_start = agent.session_start
+            review_agent._cached_system_prompt = agent._cached_system_prompt
+            # Defensive: pin session_start + session_id to the
+            # parent's so any code path that re-renders parts of
+            # the system prompt (compression, plugin hooks) still
+            # produces byte-identical output. The cached-prompt
+            # assignment above already short-circuits the normal
+            # rebuild path, but these pins guarantee parity even
+            # if a future code path bypasses the cache.
+            review_agent.session_start = agent.session_start
             review_agent.session_id = agent.session_id
             # The fork shares the parent's live session_id (pinned above for
             # prefix-cache parity). It is single-lifecycle and calls close()
@@ -812,17 +600,10 @@ def _run_review_in_thread(
                 clear_thread_tool_whitelist,
             )
 
-            # Gate the built-in memory tool on the profile's memory_enabled flag.
-            # Hardcoding ["memory", "skills"] granted the review LLM the MEMORY.md
-            # read/write tool even when a profile set memory_enabled: false,
-            # contaminating a memory-disabled profile (#54937 layer 2).
-            review_toolsets = ["skills"]
-            if review_agent._memory_enabled or review_agent._user_profile_enabled:
-                review_toolsets.insert(0, "memory")
             review_whitelist = {
                 t["function"]["name"]
                 for t in get_tool_definitions(
-                    enabled_toolsets=review_toolsets,
+                    enabled_toolsets=["memory", "skills"],
                     quiet_mode=True,
                 )
             }
@@ -834,20 +615,6 @@ def _run_review_in_thread(
                 ),
             )
             try:
-                from tools.skill_manager_tool import _reset_background_review_read_marks
-
-                _reset_background_review_read_marks()
-            except Exception:
-                pass
-
-            try:
-                # Routed to a different model -> replay a digest (cache is cold
-                # on that model anyway, so minimise cold-written tokens). Same
-                # model -> replay the full snapshot (warm cache reads).
-                _review_history = (
-                    _digest_history(messages_snapshot) if _routed
-                    else messages_snapshot
-                )
                 review_agent.run_conversation(
                     user_message=(
                         prompt
@@ -855,7 +622,7 @@ def _run_review_in_thread(
                         "management tools. Other tools will be denied "
                         "at runtime — do not attempt them."
                     ),
-                    conversation_history=_review_history,
+                    conversation_history=messages_snapshot,
                 )
             finally:
                 clear_thread_tool_whitelist()
@@ -885,29 +652,11 @@ def _run_review_in_thread(
         # the review agent inherits that history and would otherwise
         # re-surface stale "created"/"updated" messages from the prior
         # conversation as if they just happened (issue #14944).
-        #
-        # Wrapped in try/except: a buggy/legacy tool response shape
-        # (e.g. ``_change`` returned as a list instead of a dict, #59437)
-        # must NOT take down the whole review with an AttributeError,
-        # since the caller's outer except logs only "Background
-        # memory/skill review failed" and discards every successful
-        # action the fork DID complete before the crash. Coerce an
-        # exception into an empty actions list so the partial valid
-        # actions from earlier in the messages are returned instead.
-        try:
-            actions = summarize_background_review_actions(
-                review_messages,
-                messages_snapshot,
-                notification_mode=getattr(agent, "memory_notifications", "on"),
-            )
-        except Exception as e:
-            logger.warning(
-                "summarize_background_review_actions returned partial results "
-                "after exception (treating as empty); suppressing AttributeError "
-                "that previously aborted the entire review (#59437): %s",
-                e,
-            )
-            actions = []
+        actions = summarize_background_review_actions(
+            review_messages,
+            messages_snapshot,
+            notification_mode=getattr(agent, "memory_notifications", "on"),
+        )
 
         if actions:
             summary = " · ".join(dict.fromkeys(actions))
@@ -927,14 +676,16 @@ def _run_review_in_thread(
         logger.warning("Background memory/skill review failed: %s", e)
         agent._emit_auxiliary_failure("background review", e)
     finally:
-        # Safety-net cleanup for the exception path.  Normal completion already
-        # shut down inside the thread-scoped silence above.  Re-enter the
-        # thread-scoped silence here so teardown output (Honcho flush, Hindsight
-        # sync, background thread joins) stays quiet even on the exception path,
-        # without blanking other threads' streams.
+        # Safety-net cleanup for the exception path.  Normal
+        # completion already shut down inside redirect_stdout above.
+        # Re-open devnull here so any teardown output (Honcho flush,
+        # Hindsight sync, background thread joins) stays silent even
+        # on the exception path where redirect_stdout already exited.
         if review_agent is not None:
             try:
-                with thread_scoped_silence():
+                with open(os.devnull, "w", encoding="utf-8") as _fn, \
+                     contextlib.redirect_stdout(_fn), \
+                     contextlib.redirect_stderr(_fn):
                     try:
                         review_agent.shutdown_memory_provider()
                     except Exception:

agent/battery.py

@@ -1,131 +0,0 @@
-"""System-battery read-out for the CLI/TUI status bar.
-
-Reads the host battery through ``psutil`` (already a Hermes dependency) and
-exposes a compact, colour-coded label.  Everything degrades to "unavailable"
-when there is no battery (desktops, servers, VMs) or when the read fails, so
-callers can render the result unconditionally and simply show nothing.
-
-The status bar repaints often (every keystroke and on a ~1s idle refresh), so
-:func:`read_battery` memoises the last reading for a few seconds instead of
-hitting ``psutil`` on every frame.
-"""
-
-from __future__ import annotations
-
-import time
-from dataclasses import dataclass
-from typing import Optional
-
-
-@dataclass(frozen=True)
-class BatteryStatus:
-    """A single battery reading.
-
-    ``available`` is False on machines without a battery (or when the read
-    failed).  ``percent`` is clamped to 0-100.  ``plugged`` is True when on AC
-    power, False on battery, and None when the platform can't tell.
-    """
-
-    available: bool
-    percent: Optional[int] = None
-    plugged: Optional[bool] = None
-
-    @property
-    def charging(self) -> bool:
-        return bool(self.plugged)
-
-
-UNAVAILABLE = BatteryStatus(available=False)
-
-# Colour buckets, mirroring the status-bar context styles but inverted (a full
-# battery is "good", an empty one is "critical").
-CATEGORY_GOOD = "good"
-CATEGORY_WARN = "warn"
-CATEGORY_BAD = "bad"
-CATEGORY_CRITICAL = "critical"
-CATEGORY_DIM = "dim"
-
-_CACHE_TTL_SECONDS = 8.0
-_cache: Optional[tuple[float, BatteryStatus]] = None
-
-
-def _read_battery_uncached() -> BatteryStatus:
-    try:
-        import psutil
-    except Exception:
-        return UNAVAILABLE
-
-    # ``sensors_battery`` is missing on some platforms/builds of psutil.
-    reader = getattr(psutil, "sensors_battery", None)
-    if reader is None:
-        return UNAVAILABLE
-
-    try:
-        batt = reader()
-    except Exception:
-        return UNAVAILABLE
-
-    if batt is None:
-        return UNAVAILABLE
-
-    percent: Optional[int] = None
-    raw_percent = getattr(batt, "percent", None)
-    if raw_percent is not None:
-        try:
-            percent = max(0, min(100, int(round(float(raw_percent)))))
-        except (TypeError, ValueError):
-            percent = None
-
-    plugged = getattr(batt, "power_plugged", None)
-    if plugged is not None:
-        plugged = bool(plugged)
-
-    return BatteryStatus(available=True, percent=percent, plugged=plugged)
-
-
-def read_battery(use_cache: bool = True) -> BatteryStatus:
-    """Return the current battery status (cached for a few seconds)."""
-    global _cache
-    if use_cache and _cache is not None:
-        ts, cached = _cache
-        if time.monotonic() - ts < _CACHE_TTL_SECONDS:
-            return cached
-
-    status = _read_battery_uncached()
-    _cache = (time.monotonic(), status)
-    return status
-
-
-def clear_cache() -> None:
-    """Drop the memoised reading (used by tests)."""
-    global _cache
-    _cache = None
-
-
-def battery_category(status: BatteryStatus) -> str:
-    """Bucket a reading into a colour category: good/warn/bad/critical/dim."""
-    if not status.available or status.percent is None:
-        return CATEGORY_DIM
-    # On AC power the level isn't a concern — always read as healthy.
-    if status.charging:
-        return CATEGORY_GOOD
-    pct = status.percent
-    if pct <= 10:
-        return CATEGORY_CRITICAL
-    if pct <= 20:
-        return CATEGORY_BAD
-    if pct <= 50:
-        return CATEGORY_WARN
-    return CATEGORY_GOOD
-
-
-def battery_glyph(status: BatteryStatus) -> str:
-    """Return the leading glyph: a bolt while charging, else a battery."""
-    return "\u26a1" if status.charging else "\U0001f50b"  # ⚡ / 🔋
-
-
-def format_battery(status: BatteryStatus) -> str:
-    """Return a compact label like ``🔋 82%`` / ``⚡ 82%`` (empty if N/A)."""
-    if not status.available or status.percent is None:
-        return ""
-    return f"{battery_glyph(status)} {status.percent}%"

agent/bedrock_adapter.py

@@ -433,29 +433,6 @@ def _model_supports_tool_use(model_id: str) -> bool:
     return not any(pattern in model_lower for pattern in _NON_TOOL_CALLING_PATTERNS)
 
 
-# ---------------------------------------------------------------------------
-# Prompt-cache capability detection (Converse API cachePoint)
-# ---------------------------------------------------------------------------
-# Claude on Bedrock already gets prompt caching through the AnthropicBedrock
-# SDK path (see is_anthropic_bedrock_model / runtime_provider.py's dual-path
-# routing) — it never reaches build_converse_kwargs unless bearer-token auth
-# forces the Converse path (#28156). This allowlist covers the Converse API
-# itself: sending an unsupported model a cachePoint block raises a
-# ValidationException, so — like _model_supports_tool_use but inverted —
-# unknown models default to NOT receiving cache markers until confirmed.
-# Ref: https://docs.aws.amazon.com/bedrock/latest/userguide/prompt-caching.html
-_CACHE_POINT_PATTERNS = [
-    "anthropic.claude",  # bearer-token fallback path
-    "amazon.nova",
-]
-
-
-def _model_supports_prompt_cache(model_id: str) -> bool:
-    """Return True if the model accepts a Converse API cachePoint block."""
-    model_lower = model_id.lower()
-    return any(pattern in model_lower for pattern in _CACHE_POINT_PATTERNS)
-
-
 def is_anthropic_bedrock_model(model_id: str) -> bool:
     """Return True if the model is an Anthropic Claude model on Bedrock.
 
@@ -471,10 +448,7 @@ def is_anthropic_bedrock_model(model_id: str) -> bool:
     """
     model_lower = model_id.lower()
     # Strip regional prefix if present
-    for prefix in (
-        "global.", "us.", "eu.", "apac.", "ap.", "au.", "jp.",
-        "ca.", "sa.", "me.", "af.",
-    ):
+    for prefix in ("us.", "global.", "eu.", "ap.", "jp."):
         if model_lower.startswith(prefix):
             model_lower = model_lower[len(prefix):]
             break
@@ -516,26 +490,6 @@ def convert_tools_to_converse(tools: List[Dict]) -> List[Dict]:
     return result
 
 
-# Bedrock's Converse API rejects any text content block whose text is empty
-# OR whitespace-only (ValidationException: "text content blocks must contain
-# non-whitespace text"). A lone space is whitespace and is rejected too — the
-# placeholder MUST itself be non-whitespace. Ref: issue #9486.
-_EMPTY_TEXT_PLACEHOLDER = "(empty)"
-
-
-def _safe_text(text) -> str:
-    """Return ``text`` if it's non-whitespace, else a non-whitespace placeholder.
-
-    Handles None, empty string, and whitespace-only string (spaces, tabs,
-    newlines) — all of which Bedrock's Converse API rejects as text content.
-    """
-    if text is None:
-        return _EMPTY_TEXT_PLACEHOLDER
-    if not isinstance(text, str):
-        text = str(text)
-    return text if text.strip() else _EMPTY_TEXT_PLACEHOLDER
-
-
 def _convert_content_to_converse(content) -> List[Dict]:
     """Convert OpenAI message content (string or list) to Converse content blocks.
 
@@ -543,27 +497,26 @@ def _convert_content_to_converse(content) -> List[Dict]:
       - Plain text strings → [{"text": "..."}]
       - Content arrays with text/image_url parts → mixed text/image blocks
 
-    Replaces empty/whitespace-only text blocks with a non-whitespace
-    placeholder — Bedrock's Converse API rejects messages where a text
-    content block is empty or whitespace-only (ValidationException:
-    "text content blocks must contain non-whitespace text"). Ref: issue #9486.
+    Filters out empty text blocks — Bedrock's Converse API rejects messages
+    where a text content block has an empty ``text`` field (ValidationException:
+    "text content blocks must be non-empty"). Ref: issue #9486.
     """
     if content is None:
-        return [{"text": _safe_text(content)}]
+        return [{"text": " "}]
     if isinstance(content, str):
-        return [{"text": _safe_text(content)}]
+        return [{"text": content}] if content.strip() else [{"text": " "}]
     if isinstance(content, list):
         blocks = []
         for part in content:
             if isinstance(part, str):
-                blocks.append({"text": _safe_text(part)})
+                blocks.append({"text": part})
                 continue
             if not isinstance(part, dict):
                 continue
             part_type = part.get("type", "")
             if part_type == "text":
                 text = part.get("text", "")
-                blocks.append({"text": _safe_text(text)})
+                blocks.append({"text": text if text else " "})
             elif part_type == "image_url":
                 image_url = part.get("image_url", {})
                 url = image_url.get("url", "") if isinstance(image_url, dict) else ""
@@ -575,27 +528,18 @@ def _convert_content_to_converse(content) -> List[Dict]:
                         mime_part = header[5:].split(";")[0]
                         if mime_part:
                             media_type = mime_part
-                    # Decode base64 to raw bytes — boto3 re-encodes at the
-                    # wire layer, so passing the base64 string directly
-                    # results in double-encoding and Bedrock rejects it with
-                    # "Failed to sanitize image".  Ref: #33317.
-                    import base64
-                    try:
-                        raw_bytes = base64.b64decode(data)
-                    except Exception:
-                        raw_bytes = data.encode("utf-8")
                     blocks.append({
                         "image": {
                             "format": media_type.split("/")[-1] if "/" in media_type else "jpeg",
-                            "source": {"bytes": raw_bytes},
+                            "source": {"bytes": data},
                         }
                     })
                 else:
                     # Remote URL — Converse doesn't support URLs directly,
                     # include as text reference for the model.
                     blocks.append({"text": f"[Image: {url}]"})
-        return blocks if blocks else [{"text": _EMPTY_TEXT_PLACEHOLDER}]
-    return [{"text": _safe_text(content)}]
+        return blocks if blocks else [{"text": " "}]
+    return [{"text": str(content)}]
 
 
 def convert_messages_to_converse(
@@ -625,18 +569,14 @@ def convert_messages_to_converse(
         content = msg.get("content")
 
         if role == "system":
-            # System messages become the system prompt. Blank/whitespace-only
-            # parts are dropped entirely (not placeholder-filled) since a
-            # system prompt made up of only placeholder text is meaningless.
+            # System messages become the system prompt
             if isinstance(content, str) and content.strip():
                 system_blocks.append({"text": content})
             elif isinstance(content, list):
                 for part in content:
                     if isinstance(part, dict) and part.get("type") == "text":
-                        text = part.get("text", "")
-                        if isinstance(text, str) and text.strip():
-                            system_blocks.append({"text": text})
-                    elif isinstance(part, str) and part.strip():
+                        system_blocks.append({"text": part.get("text", "")})
+                    elif isinstance(part, str):
                         system_blocks.append({"text": part})
             continue
 
@@ -647,7 +587,7 @@ def convert_messages_to_converse(
             tool_result_block = {
                 "toolResult": {
                     "toolUseId": tool_call_id,
-                    "content": [{"text": _safe_text(result_content)}],
+                    "content": [{"text": result_content}],
                 }
             }
             # In Converse, tool results go in a "user" role message
@@ -686,7 +626,7 @@ def convert_messages_to_converse(
                 })
 
             if not content_blocks:
-                content_blocks = [{"text": _EMPTY_TEXT_PLACEHOLDER}]
+                content_blocks = [{"text": " "}]
 
             # Merge with previous assistant message if needed (strict alternation)
             if converse_msgs and converse_msgs[-1]["role"] == "assistant":
@@ -712,11 +652,11 @@ def convert_messages_to_converse(
 
     # Converse requires the first message to be from the user
     if converse_msgs and converse_msgs[0]["role"] != "user":
-        converse_msgs.insert(0, {"role": "user", "content": [{"text": _EMPTY_TEXT_PLACEHOLDER}]})
+        converse_msgs.insert(0, {"role": "user", "content": [{"text": " "}]})
 
     # Converse requires the last message to be from the user
     if converse_msgs and converse_msgs[-1]["role"] != "user":
-        converse_msgs.append({"role": "user", "content": [{"text": _EMPTY_TEXT_PLACEHOLDER}]})
+        converse_msgs.append({"role": "user", "content": [{"text": " "}]})
 
     return (system_blocks if system_blocks else None, converse_msgs)
 
@@ -787,22 +727,14 @@ def normalize_converse_response(response: Dict) -> SimpleNamespace:
         reasoning_content="\n\n".join(reasoning_parts) if reasoning_parts else None,
     )
 
-    # Build usage stats. Converse's inputTokens excludes cache read/write
-    # tokens (unlike OpenAI's prompt_tokens, which includes them) — restore
-    # the OpenAI-style "total includes cache" convention here so downstream
-    # normalize_usage() can subtract them back out consistently, and surface
-    # the Anthropic-named fields it already falls back to for cache reads.
+    # Build usage stats
     usage_data = response.get("usage", {})
-    input_tokens = usage_data.get("inputTokens", 0)
-    cache_read_tokens = usage_data.get("cacheReadInputTokens", 0)
-    cache_write_tokens = usage_data.get("cacheWriteInputTokens", 0)
-    output_tokens = usage_data.get("outputTokens", 0)
     usage = SimpleNamespace(
-        prompt_tokens=input_tokens + cache_read_tokens + cache_write_tokens,
-        completion_tokens=output_tokens,
-        total_tokens=input_tokens + cache_read_tokens + cache_write_tokens + output_tokens,
-        cache_read_input_tokens=cache_read_tokens,
-        cache_creation_input_tokens=cache_write_tokens,
+        prompt_tokens=usage_data.get("inputTokens", 0),
+        completion_tokens=usage_data.get("outputTokens", 0),
+        total_tokens=(
+            usage_data.get("inputTokens", 0) + usage_data.get("outputTokens", 0)
+        ),
     )
 
     finish_reason = _converse_stop_reason_to_openai(stop_reason)
@@ -848,7 +780,6 @@ def stream_converse_with_callbacks(
     on_tool_start=None,
     on_reasoning_delta=None,
     on_interrupt_check=None,
-    on_event=None,
 ) -> SimpleNamespace:
     """Process a Bedrock ConverseStream event stream with real-time callbacks.
 
@@ -868,12 +799,6 @@ def stream_converse_with_callbacks(
             on supported models (Claude 4.6+).
         on_interrupt_check: Called on each event. Should return True if the
             agent has been interrupted and streaming should stop.
-        on_event: Called once at the top of the loop body for EVERY yielded
-            Bedrock event (text/tool-input/reasoning/metadata deltas alike),
-            before any branching. Provides a wire-level liveness signal so an
-            external watchdog can distinguish "still receiving events" from
-            "stream wedged with no data". Errors raised by the callback are
-            swallowed so a liveness hook can never abort the stream.
 
     Returns:
         An OpenAI-compatible SimpleNamespace response, identical in shape to
@@ -889,15 +814,6 @@ def stream_converse_with_callbacks(
     usage_data: Dict[str, int] = {}
 
     for event in event_stream.get("stream", []):
-        # Wire-level liveness signal: fire on EVERY yielded event (text, tool
-        # input, reasoning, metadata) before branching so an external watchdog
-        # can tell a still-flowing stream from a wedged one. Best-effort — a
-        # liveness callback must never be able to abort the stream.
-        if on_event is not None:
-            try:
-                on_event()
-            except Exception:
-                pass
         # Check for interrupt
         if on_interrupt_check and on_interrupt_check():
             break
@@ -967,8 +883,6 @@ def stream_converse_with_callbacks(
             usage_data = {
                 "inputTokens": meta_usage.get("inputTokens", 0),
                 "outputTokens": meta_usage.get("outputTokens", 0),
-                "cacheReadInputTokens": meta_usage.get("cacheReadInputTokens", 0),
-                "cacheWriteInputTokens": meta_usage.get("cacheWriteInputTokens", 0),
             }
 
     # Flush remaining text
@@ -982,16 +896,12 @@ def stream_converse_with_callbacks(
         reasoning_content="\n\n".join(reasoning_parts) if reasoning_parts else None,
     )
 
-    input_tokens = usage_data.get("inputTokens", 0)
-    cache_read_tokens = usage_data.get("cacheReadInputTokens", 0)
-    cache_write_tokens = usage_data.get("cacheWriteInputTokens", 0)
-    output_tokens = usage_data.get("outputTokens", 0)
     usage = SimpleNamespace(
-        prompt_tokens=input_tokens + cache_read_tokens + cache_write_tokens,
-        completion_tokens=output_tokens,
-        total_tokens=input_tokens + cache_read_tokens + cache_write_tokens + output_tokens,
-        cache_read_input_tokens=cache_read_tokens,
-        cache_creation_input_tokens=cache_write_tokens,
+        prompt_tokens=usage_data.get("inputTokens", 0),
+        completion_tokens=usage_data.get("outputTokens", 0),
+        total_tokens=(
+            usage_data.get("inputTokens", 0) + usage_data.get("outputTokens", 0)
+        ),
     )
 
     finish_reason = _converse_stop_reason_to_openai(stop_reason)
@@ -1030,7 +940,6 @@ def build_converse_kwargs(
     Converts OpenAI-format inputs to Converse API parameters.
     """
     system_prompt, converse_messages = convert_messages_to_converse(messages)
-    cache_enabled = _model_supports_prompt_cache(model)
 
     kwargs: Dict[str, Any] = {
         "modelId": model,
@@ -1041,8 +950,6 @@ def build_converse_kwargs(
     }
 
     if system_prompt:
-        if cache_enabled:
-            system_prompt = system_prompt + [{"cachePoint": {"type": "default"}}]
         kwargs["system"] = system_prompt
 
     from agent.anthropic_adapter import _forbids_sampling_params
@@ -1066,8 +973,6 @@ def build_converse_kwargs(
             # Strip tools for known non-tool-calling models and warn the user.
             # Ref: PR #7920 feedback from @ptlally, pattern from PR #4346.
             if _model_supports_tool_use(model):
-                if cache_enabled:
-                    converse_tools = converse_tools + [{"cachePoint": {"type": "default"}}]
                 kwargs["toolConfig"] = {"tools": converse_tools}
             else:
                 logger.warning(
@@ -1075,14 +980,6 @@ def build_converse_kwargs(
                     "The agent will operate in text-only mode.", model
                 )
 
-    if cache_enabled and len(converse_messages) >= 2:
-        # Checkpoint everything up to (not including) the newest turn, so the
-        # marker survives unchanged across requests as only the tail grows —
-        # mirroring the Anthropic system_and_3 strategy in prompt_caching.py.
-        content = converse_messages[-2].get("content")
-        if isinstance(content, list) and content:
-            content.append({"cachePoint": {"type": "default"}})
-
     if guardrail_config:
         kwargs["guardrailConfig"] = guardrail_config
 
@@ -1399,24 +1296,9 @@ def classify_bedrock_error(error_message: str) -> str:
 # detection is unavailable.
 
 BEDROCK_CONTEXT_LENGTHS: Dict[str, int] = {
-    # Anthropic Claude models on Bedrock.
-    # Context windows per Anthropic's official models comparison
-    # (https://platform.claude.com/docs/en/about-claude/models/overview).
-    # Fable / Sonnet 5 / Opus 4.8 / 4.7 / 4.6 / Sonnet 4.6 have 1M generally
-    # available (no beta header required as of April 2026). Sonnet 4.5 and
-    # Sonnet 4 had their `context-1m-2025-08-07` beta retired on
-    # April 30, 2026, so they are standard 200K; Haiku 4.5 is 200K.
-    # These 1M entries must match agent/model_metadata.py
-    # DEFAULT_CONTEXT_LENGTHS or the agent compresses context prematurely.
-    # Keys are matched by longest-substring, so the versioned 4-6/4-7/4-8
-    # entries win over the generic "anthropic.claude-opus-4" fallback.
-    "anthropic.claude-fable-5":      1_000_000,
-    "anthropic.claude-fable":        1_000_000,
-    "anthropic.claude-sonnet-5":     1_000_000,
-    "anthropic.claude-opus-4-8":     1_000_000,
-    "anthropic.claude-opus-4-7":     1_000_000,
-    "anthropic.claude-opus-4-6":     1_000_000,
-    "anthropic.claude-sonnet-4-6":   1_000_000,
+    # Anthropic Claude models on Bedrock
+    "anthropic.claude-opus-4-6":     200_000,
+    "anthropic.claude-sonnet-4-6":   200_000,
     "anthropic.claude-sonnet-4-5":   200_000,
     "anthropic.claude-haiku-4-5":    200_000,
     "anthropic.claude-opus-4":       200_000,
@@ -1443,22 +1325,9 @@ BEDROCK_CONTEXT_LENGTHS: Dict[str, int] = {
 # Default for unknown Bedrock models
 BEDROCK_DEFAULT_CONTEXT_LENGTH = 128_000
 
-# Probe tiers (in tokens).  We send a request padded just past each tier and
-# read the real window from Bedrock's length-validation error.  Two reasons
-# this is tiered rather than one giant request:
-#   1. A wildly oversized payload (e.g. 5M tokens) makes Bedrock return an
-#      opaque InternalServerException after retries instead of a clean
-#      ValidationException — so we must stay within a sane overage.
-#   2. Stepping up lets us discover larger windows (2M+) without over-padding
-#      smaller ones.
-# Each tier value is the *padding target*; the error reports the true maximum,
-# which is what we actually return.
-_BEDROCK_PROBE_TIERS = (1_300_000, 2_200_000)
-_WORDS_PER_TOKEN = 0.9  # conservative: ensures the padded prompt clears the tier
 
-
-def _static_bedrock_context_length(model_id: str) -> int:
-    """Longest-substring-match lookup against the static fallback table.
+def get_bedrock_context_length(model_id: str) -> int:
+    """Look up the context window size for a Bedrock model.
 
     Uses substring matching so versioned IDs like
     ``anthropic.claude-sonnet-4-6-20250514-v1:0`` resolve correctly.
@@ -1471,103 +1340,3 @@ def _static_bedrock_context_length(model_id: str) -> int:
             best_key = key
             best_val = val
     return best_val
-
-
-def probe_bedrock_context_length(model_id: str, region: str) -> Optional[int]:
-    """Discover a Bedrock model's real context window by provoking a length error.
-
-    Bedrock does not expose the context window via any metadata API
-    (``get-foundation-model`` omits it, ``Converse`` metrics omit it,
-    ``CountTokens`` is unsupported on several models).  The only authoritative
-    source is the ``ValidationException`` raised when a prompt exceeds the
-    window:
-
-        "The model returned the following errors: prompt is too long:
-         1300032 tokens > 1000000 maximum"
-
-    Length validation happens *before* inference, so an oversized request is
-    rejected immediately and cheaply — no tokens are generated and no input is
-    actually processed.  We pad a request just past each tier in
-    ``_BEDROCK_PROBE_TIERS`` and parse the reported ``maximum``.  Tiers exist
-    because (a) a *wildly* oversized payload makes Bedrock fail with an opaque
-    InternalServerException instead of a clean length error, and (b) stepping
-    up discovers larger windows without over-padding smaller ones.
-
-    Returns the detected window, or ``None`` if the probe could not run
-    (missing credentials, network error, or no parseable limit) so the caller
-    can fall back to the static table.
-    """
-    try:
-        from agent.model_metadata import parse_context_limit_from_error
-    except ImportError:  # pragma: no cover — same package
-        return None
-
-    try:
-        client = _get_bedrock_runtime_client(region)
-    except Exception as exc:  # boto3 missing / credential resolution failure
-        logger.debug("Bedrock context probe skipped for %s: %s", model_id, exc)
-        return None
-
-    last_error = ""
-    for tier_tokens in _BEDROCK_PROBE_TIERS:
-        pad_words = int(tier_tokens / _WORDS_PER_TOKEN)
-        oversized = "data " * pad_words
-        try:
-            client.converse(
-                modelId=model_id,
-                messages=[{"role": "user", "content": [{"text": oversized}]}],
-                inferenceConfig={"maxTokens": 8},
-            )
-            # Accepted a prompt this large → the window is at least this tier.
-            # Returning the tier as a lower bound is safe and avoids inventing
-            # a number we can't confirm.
-            logger.debug(
-                "Bedrock context probe for %s accepted ~%s-token prompt; "
-                "window is at least that", model_id, f"{tier_tokens:,}",
-            )
-            return tier_tokens
-        except Exception as exc:
-            msg = str(exc)
-            last_error = msg
-            limit = parse_context_limit_from_error(msg)
-            if limit and limit >= 1024:
-                logger.info(
-                    "Probed Bedrock context window for %s: %s tokens",
-                    model_id, f"{limit:,}",
-                )
-                return limit
-            # No parseable limit at this tier (opaque server error, auth,
-            # throttle).  Try the next, smaller-overage strategy is N/A here —
-            # tiers ascend — so just continue; if all fail we return None.
-            continue
-
-    logger.debug(
-        "Bedrock context probe for %s returned no parseable limit: %s",
-        model_id, last_error[:200],
-    )
-    return None
-
-
-def get_bedrock_context_length(model_id: str, region: str = "", probe: bool = True) -> int:
-    """Resolve the context window for a Bedrock model.
-
-    Resolution order:
-      1. Live probe against Bedrock (authoritative; cached by the caller).
-      2. Static fallback table (longest-substring match).
-      3. Conservative default.
-
-    The static table is intentionally a *fallback*, not the primary source:
-    AWS ships new model versions (opus-4-7, opus-4-8, ...) faster than the
-    table can track, and a stale entry silently caps the window (e.g. a
-    1M-token Opus pinned to 200K via an ``opus-4`` substring match).  The
-    probe asks Bedrock directly so every model — current or future — gets its
-    real window with no table maintenance.
-
-    ``probe=False`` (or an empty ``region``) skips the network call and uses
-    the static table only — used by pure-offline/display code paths.
-    """
-    if probe and region:
-        probed = probe_bedrock_context_length(model_id, region)
-        if probed:
-            return probed
-    return _static_bedrock_context_length(model_id)

agent/billing_links.py

@@ -1,124 +0,0 @@
-"""Provider-agnostic billing/credit recovery links.
-
-Maps a billing-classified failure onto a recovery link + label. *Detection*
-is not done here — that is :mod:`agent.error_classifier`
-(``FailoverReason.billing``), the single source of truth for "credit wall vs.
-rate limit / auth / transport". The resulting :class:`BillingBlock` rides the
-turn result and the gateway ``message.complete`` event so every surface (CLI,
-TUI, desktop) renders one structured signal instead of re-parsing error text.
-"""
-
-from __future__ import annotations
-
-from dataclasses import asdict, dataclass
-from typing import Optional
-
-from utils import base_url_host_matches
-
-
-@dataclass
-class BillingBlock:
-    """Structured billing-wall descriptor shared across every surface.
-
-    ``is_nous`` is the routing bit: Nous has a first-class in-app billing surface
-    (desktop Settings → Billing, TUI/CLI ``/topup``), so surfaces prefer that over
-    ``billing_url``; third-party providers have no in-app flow, so ``billing_url``
-    is the deep link the user actually needs.
-    """
-
-    provider: str
-    provider_label: str
-    model: str
-    billing_url: Optional[str]
-    is_nous: bool
-    message: str
-
-    def to_dict(self) -> dict:
-        return asdict(self)
-
-
-@dataclass(frozen=True)
-class _Provider:
-    label: str
-    url: str
-    slugs: tuple[str, ...]
-    hosts: tuple[str, ...] = ()
-
-
-# Single source of truth: internal slug(s) + base_url host(s) → billing page.
-# Curated "add credits / manage billing" landing pages, not marketing homes.
-# Hosts back the OpenAI-compatible fallback where the slug is a generic bucket
-# (e.g. "openai_compatible") but base_url reveals the real upstream. An unknown
-# provider degrades to a readable label with no invented URL.
-_PROVIDERS: tuple[_Provider, ...] = (
-    _Provider("OpenAI", "https://platform.openai.com/settings/organization/billing", ("openai",), ("api.openai.com",)),
-    _Provider("Anthropic", "https://console.anthropic.com/settings/billing", ("anthropic",), ("api.anthropic.com",)),
-    _Provider("OpenRouter", "https://openrouter.ai/settings/credits", ("openrouter",), ("openrouter.ai",)),
-    _Provider("xAI", "https://console.x.ai/team/default/billing", ("xai", "xai-oauth"), ("api.x.ai",)),
-    _Provider("DeepSeek", "https://platform.deepseek.com/top_up", ("deepseek",), ("api.deepseek.com",)),
-    _Provider("Groq", "https://console.groq.com/settings/billing", ("groq",), ("api.groq.com",)),
-    _Provider("Mistral", "https://console.mistral.ai/billing", ("mistral",), ("api.mistral.ai",)),
-    _Provider("Together AI", "https://api.together.ai/settings/billing", ("together",), ("api.together.ai", "api.together.xyz")),
-    _Provider("Fireworks AI", "https://fireworks.ai/account/billing", ("fireworks",), ("fireworks.ai",)),
-    _Provider("Perplexity", "https://www.perplexity.ai/settings/api", ("perplexity",), ("perplexity.ai",)),
-    _Provider("Google AI", "https://aistudio.google.com/app/billing", ("google", "gemini"), ("generativelanguage.googleapis.com",)),
-    _Provider("Cohere", "https://dashboard.cohere.com/billing", ("cohere",)),
-    _Provider("Moonshot AI", "https://platform.moonshot.ai/console/pay", ("moonshot",)),
-    _Provider("NVIDIA", "https://build.nvidia.com/settings/billing", ("nvidia",)),
-)
-
-_BY_SLUG: dict[str, _Provider] = {slug: p for p in _PROVIDERS for slug in p.slugs}
-
-
-def is_nous_inference_route(provider: str, base_url: str) -> bool:
-    """True when the failing route is the Nous-managed inference gateway."""
-    if (provider or "").strip().lower() == "nous":
-        return True
-    return base_url_host_matches(str(base_url or ""), "inference-api.nousresearch.com")
-
-
-def _nous_billing_url() -> Optional[str]:
-    """Best-effort Nous portal billing URL (text-surface fallback; Nous prefers the in-app flow)."""
-    try:
-        from hermes_cli.nous_account import nous_portal_billing_url
-
-        return nous_portal_billing_url(None)
-    except Exception:
-        return "https://portal.nousresearch.com/billing"
-
-
-def _resolve_provider_link(slug: str, base_url: str) -> tuple[str, Optional[str]]:
-    """Resolve ``(label, url)``: exact slug → base_url host → readable-label fallback."""
-    hit = _BY_SLUG.get(slug)
-    if hit:
-        return hit.label, hit.url
-
-    base = str(base_url or "")
-    for p in _PROVIDERS:
-        if any(base_url_host_matches(base, host) for host in p.hosts):
-            return p.label, p.url
-
-    return slug.replace("_", " ").replace("-", " ").strip().title() or "your provider", None
-
-
-def build_billing_block(
-    *,
-    provider: str,
-    base_url: str,
-    model: str,
-    message: str = "",
-) -> BillingBlock:
-    """Build the billing descriptor for a billing-classified failure.
-
-    ``message`` is the guidance already assembled by the agent loop
-    (:func:`agent.conversation_loop._billing_or_entitlement_message`), carried
-    through unchanged so every surface shows identical copy.
-    """
-    slug = (provider or "").strip().lower()
-    model = (model or "").strip()
-
-    if is_nous_inference_route(slug, base_url):
-        return BillingBlock(slug or "nous", "Nous Portal", model, _nous_billing_url(), True, message or "")
-
-    label, url = _resolve_provider_link(slug, base_url)
-    return BillingBlock(slug, label, model, url, False, message or "")

agent/billing_usage.py

@@ -1,323 +0,0 @@
-"""Shared dollar-denominated usage model for the billing/subscription surfaces.
-
-The single source of truth behind the ``/usage`` and ``/subscription`` usage
-bars (TUI + CLI). User feedback (Jun 2026): the terminal surfaces show
-**dollars**, never "credits", and every usage bar must make the monthly
-subscription allowance and separately-purchased top-up dollars distinctly
-visible.
-
-Data source: the NAS account-info fetch (``NousPortalAccountInfo``), whose
-``paid_service_access_info`` carries the three dollar magnitudes we render
-(despite the legacy ``*_credits`` field names, these are USD floats):
-
-  - ``subscription_credits_remaining``  -> plan dollars left this month
-  - ``purchased_credits_remaining``     -> top-up dollars left (rolls over)
-  - ``total_usable_credits``            -> total spendable
-
-plus ``subscription.monthly_credits`` (the plan's monthly $ allowance, the
-denominator for the "% used" plan bar) and ``current_period_end`` (renewal).
-
-Design: two SEPARATE bars (decided with the user) rather than one crammed
-three-segment bar — at terminal widths three same-glyph density segments are
-unreadable. The plan bar is "spent vs allowance this month" (carries % used);
-the top-up bar is "money you bought, doesn't expire". Each gets full
-resolution and a single fill glyph, so the bar is never ambiguous and never
-relies on color.
-
-Fail-open everywhere: any missing/non-finite field degrades to fewer bars or a
-magnitudes-only view; a logged-out / unreachable portal yields
-``available=False`` and the surface shows nothing.
-"""
-
-from __future__ import annotations
-
-import logging
-import math
-import os
-from dataclasses import dataclass, field
-from typing import Any, Optional
-
-logger = logging.getLogger(__name__)
-
-# Below this TOTAL spendable ($), a paid account is flagged "low" — the alert
-# state that nudges top-up/upgrade before a mid-run cutoff. Product threshold
-# (user feedback): "any amount below $5 should be an alert status."
-LOW_BALANCE_THRESHOLD_USD = 5.0
-
-
-def _finite(value: Any) -> Optional[float]:
-    """Return value as a float iff it's a real finite number (not bool/NaN/Inf)."""
-    if isinstance(value, bool) or not isinstance(value, (int, float)):
-        return None
-    f = float(value)
-    return f if math.isfinite(f) else None
-
-
-def _fmt_usd(value: Optional[float]) -> str:
-    """``$X.YY`` for display. ``None`` -> ``$0.00`` (callers gate on presence)."""
-    return f"${(value or 0.0):,.2f}"
-
-
-def format_renews(value: Optional[str]) -> Optional[str]:
-    """Format an ISO date/timestamp as a human date, e.g. ``Jul 24, 2026``.
-
-    Accepts ``2026-07-24``, ``2026-07-24T11:05:01.000Z``, etc. Returns the raw
-    string unchanged if it can't be parsed (never raises), and ``None`` for
-    empty input.
-    """
-    if not value:
-        return None
-    from datetime import datetime
-
-    text = str(value).strip()
-    if not text:
-        return None
-    iso = text[:-1] + "+00:00" if text.endswith("Z") else text
-    try:
-        dt = datetime.fromisoformat(iso)
-    except ValueError:
-        # Fall back to a bare date prefix (YYYY-MM-DD) if present.
-        try:
-            dt = datetime.strptime(text[:10], "%Y-%m-%d")
-        except ValueError:
-            return text
-    # %-d isn't portable to Windows; build the day without a leading zero.
-    return f"{dt.strftime('%b')} {dt.day}, {dt.year}"
-
-
-@dataclass(frozen=True)
-class UsageBar:
-    """One full-resolution bar: ``spent`` of ``total``, plus a remaining figure.
-
-    ``kind`` is ``"plan"`` (monthly allowance, shows % used) or ``"topup"``
-    (purchased dollars, no denominator — ``spent`` is 0 and ``total`` ==
-    ``remaining`` so it renders as a full bar of available balance).
-    """
-
-    kind: str  # "plan" | "topup"
-    remaining_usd: float
-    total_usd: float
-    spent_usd: float = 0.0
-
-    @property
-    def pct_used(self) -> Optional[int]:
-        if self.kind != "plan" or self.total_usd <= 0:
-            return None
-        return max(0, min(100, round(self.spent_usd / self.total_usd * 100)))
-
-    @property
-    def fill_fraction(self) -> float:
-        """Fraction of the bar that should read as 'remaining' (filled)."""
-        if self.total_usd <= 0:
-            return 0.0
-        return max(0.0, min(1.0, self.remaining_usd / self.total_usd))
-
-
-@dataclass(frozen=True)
-class UsageModel:
-    """Surface-agnostic dollar usage model shared by /usage and /subscription.
-
-    ``status`` classifies the account for copy selection:
-      - ``"free"``     : no paid access / no subscription (free models only)
-      - ``"low"``      : paid, but total spendable < $5 (ALERT)
-      - ``"healthy"``  : paid, total spendable >= $5
-      - ``"depleted"`` : paid access lost (balance exhausted)
-    """
-
-    available: bool
-    status: str = "free"
-    plan_name: Optional[str] = None
-    renews_at: Optional[str] = None
-    renews_display: Optional[str] = None
-    subscription_remaining_usd: Optional[float] = None
-    topup_remaining_usd: Optional[float] = None
-    total_spendable_usd: Optional[float] = None
-    plan_bar: Optional[UsageBar] = None
-    topup_bar: Optional[UsageBar] = None
-
-    @property
-    def has_topup(self) -> bool:
-        return bool(self.topup_remaining_usd and self.topup_remaining_usd > 0)
-
-
-def usage_model_from_account(account_info: Any) -> UsageModel:
-    """Build a :class:`UsageModel` from a ``NousPortalAccountInfo``. Fail-open.
-
-    Returns ``UsageModel(available=False)`` when there's no usable account info
-    (logged out, no entitlement block). Never raises.
-    """
-    try:
-        if account_info is None or not getattr(account_info, "logged_in", False):
-            return UsageModel(available=False)
-
-        access = getattr(account_info, "paid_service_access_info", None)
-        sub = getattr(account_info, "subscription", None)
-        paid = getattr(account_info, "paid_service_access", None)
-
-        sub_remaining = _finite(getattr(access, "subscription_credits_remaining", None)) if access else None
-        topup_remaining = _finite(getattr(access, "purchased_credits_remaining", None)) if access else None
-        total_usable = _finite(getattr(access, "total_usable_credits", None)) if access else None
-
-        plan_name = getattr(sub, "plan", None) if sub is not None else None
-        renews_at = getattr(sub, "current_period_end", None) if sub is not None else None
-        monthly = _finite(getattr(sub, "monthly_credits", None)) if sub is not None else None
-
-        has_subscription = bool(plan_name) or (monthly is not None and monthly > 0)
-
-        # Total spendable: prefer the server's total; else sum the parts we have.
-        if total_usable is not None:
-            total_spendable = total_usable
-        else:
-            parts = [v for v in (sub_remaining, topup_remaining) if v is not None]
-            total_spendable = sum(parts) if parts else None
-
-        # Status classification.
-        if paid is False:
-            status = "depleted"
-        elif not has_subscription and not (topup_remaining and topup_remaining > 0):
-            # No plan and no purchased balance -> free-models-only.
-            status = "free"
-        elif total_spendable is not None and total_spendable < LOW_BALANCE_THRESHOLD_USD:
-            status = "low"
-        else:
-            status = "healthy"
-
-        # Plan bar — only with a positive monthly allowance AND a remaining we
-        # can place on it. spent = cap - remaining, clamped (a debt/over-cap
-        # balance reads as fully spent rather than a nonsensical negative).
-        plan_bar: Optional[UsageBar] = None
-        if monthly is not None and monthly > 0 and sub_remaining is not None:
-            remaining = max(0.0, min(monthly, sub_remaining))
-            plan_bar = UsageBar(
-                kind="plan",
-                remaining_usd=remaining,
-                total_usd=monthly,
-                spent_usd=max(0.0, monthly - sub_remaining),
-            )
-
-        # Top-up bar — only when there are purchased dollars to show. No
-        # denominator (top-up has no monthly cap), so it renders full = balance.
-        topup_bar: Optional[UsageBar] = None
-        if topup_remaining is not None and topup_remaining > 0:
-            topup_bar = UsageBar(
-                kind="topup",
-                remaining_usd=topup_remaining,
-                total_usd=topup_remaining,
-                spent_usd=0.0,
-            )
-
-        return UsageModel(
-            available=True,
-            status=status,
-            plan_name=plan_name,
-            renews_at=renews_at,
-            renews_display=format_renews(renews_at),
-            subscription_remaining_usd=sub_remaining,
-            topup_remaining_usd=topup_remaining,
-            total_spendable_usd=total_spendable,
-            plan_bar=plan_bar,
-            topup_bar=topup_bar,
-        )
-    except Exception:
-        logger.debug("usage ▸ model build failed (fail-open)", exc_info=True)
-        return UsageModel(available=False)
-
-
-def build_usage_model(*, timeout: float = 10.0) -> UsageModel:
-    """Fetch account-info and build the shared usage model. Fail-open.
-
-    Dev override: ``HERMES_DEV_CREDITS_FIXTURE`` short-circuits to a fixture so
-    every usage state is testable without a live account (mirrors the existing
-    ``/usage`` credits-block fixture path).
-    """
-    fixture = _dev_fixture_usage_model()
-    if fixture is not None:
-        return fixture
-
-    try:
-        from hermes_cli.auth import get_provider_auth_state
-
-        tok = (get_provider_auth_state("nous") or {}).get("access_token")
-        if not (isinstance(tok, str) and tok.strip()):
-            return UsageModel(available=False)
-    except Exception:
-        return UsageModel(available=False)
-
-    try:
-        import concurrent.futures
-
-        from hermes_cli.nous_account import get_nous_portal_account_info
-
-        with concurrent.futures.ThreadPoolExecutor(max_workers=1) as pool:
-            account = pool.submit(get_nous_portal_account_info, force_fresh=True).result(timeout=timeout)
-        return usage_model_from_account(account)
-    except Exception:
-        logger.debug("usage ▸ portal fetch failed (fail-open)", exc_info=True)
-        return UsageModel(available=False)
-
-
-# =============================================================================
-# Dev fixtures (throwaway scaffolding — env-var driven, no live portal)
-# =============================================================================
-
-
-def _dev_fixture_usage_model() -> Optional[UsageModel]:
-    """Map ``HERMES_DEV_CREDITS_FIXTURE`` to a usage model for offline UX work.
-
-    Recognized names: ``free | healthy | low | topup | depleted``. Returns
-    ``None`` when the env var is unset (real portal path runs).
-    """
-    name = (os.getenv("HERMES_DEV_CREDITS_FIXTURE") or "").strip().lower()
-    if not name:
-        return None
-
-    if name == "free":
-        return UsageModel(available=True, status="free", plan_name=None)
-
-    if name in ("healthy", "mid"):
-        return UsageModel(
-            available=True,
-            status="healthy",
-            plan_name="Plus",
-            renews_at="2026-07-01",
-            subscription_remaining_usd=14.0,
-            total_spendable_usd=14.0,
-            plan_bar=UsageBar(kind="plan", remaining_usd=14.0, total_usd=20.0, spent_usd=6.0),
-        )
-
-    if name in ("topup", "top-up"):
-        return UsageModel(
-            available=True,
-            status="healthy",
-            plan_name="Plus",
-            renews_at="2026-07-01",
-            subscription_remaining_usd=14.0,
-            topup_remaining_usd=12.0,
-            total_spendable_usd=26.0,
-            plan_bar=UsageBar(kind="plan", remaining_usd=14.0, total_usd=20.0, spent_usd=6.0),
-            topup_bar=UsageBar(kind="topup", remaining_usd=12.0, total_usd=12.0, spent_usd=0.0),
-        )
-
-    if name == "low":
-        return UsageModel(
-            available=True,
-            status="low",
-            plan_name="Plus",
-            renews_at="2026-07-01",
-            subscription_remaining_usd=3.4,
-            total_spendable_usd=3.4,
-            plan_bar=UsageBar(kind="plan", remaining_usd=3.4, total_usd=20.0, spent_usd=16.6),
-        )
-
-    if name == "depleted":
-        return UsageModel(
-            available=True,
-            status="depleted",
-            plan_name="Plus",
-            renews_at="2026-07-01",
-            subscription_remaining_usd=0.0,
-            total_spendable_usd=0.0,
-            plan_bar=UsageBar(kind="plan", remaining_usd=0.0, total_usd=20.0, spent_usd=20.0),
-        )
-
-    return None

agent/billing_view.py

@@ -1,4 +1,4 @@
-"""Surface-agnostic core for the Phase 2b Remote Spending screens.
+"""Surface-agnostic core for the Phase 2b terminal-billing screens.
 
 One fetch/parse per concern, consumed identically by the CLI handler
 (``cli.py::_show_billing``), the TUI JSON-RPC methods
@@ -15,7 +15,6 @@ We keep them as :class:`decimal.Decimal` end-to-end and only format for display.
 from __future__ import annotations
 
 import logging
-import os
 import uuid
 from dataclasses import dataclass, field
 from decimal import Decimal, InvalidOperation
@@ -65,47 +64,15 @@ def format_money(value: Optional[Decimal]) -> str:
 # =============================================================================
 
 
-# resolvedVia → the human answer to "why THIS card?". Keys are the server's card
-# resolution rungs (NAS card-on-file ladder); absent/unknown rungs render no label
-# so the display degrades cleanly on servers that don't send resolvedVia yet.
-_CARD_PROVENANCE_LABELS = {
-    "subPin": "the card on your subscription",
-    "customerDefault": "your default card saved on the portal",
-    "autoRefill": "your auto-reload card",
-}
-
-
 @dataclass(frozen=True)
 class CardInfo:
     brand: str
     last4: str
-    # NAS card-on-file field (post card-resolver): which ladder rung found the
-    # card. Defaults off so pre-resolver payloads parse unchanged.
-    resolved_via: Optional[str] = None
 
     @property
     def masked(self) -> str:
-        # A Link payment method has no card number (last4 = "") — render the
-        # brand alone, not "Link ····".
-        if not self.last4:
-            return self.brand
         return f"{self.brand} ····{self.last4}"
 
-    @property
-    def provenance(self) -> Optional[str]:
-        """Human label for why this card was picked, or None (unknown rung /
-        server too old to say)."""
-        if self.resolved_via is None:
-            return None
-        return _CARD_PROVENANCE_LABELS.get(self.resolved_via)
-
-    @property
-    def display(self) -> str:
-        """The one-line card display: ``Visa ····4242 — the card on your
-        subscription`` (or just the masked card when provenance is unknown)."""
-        label = self.provenance
-        return f"{self.masked} — {label}" if label else self.masked
-
 
 @dataclass(frozen=True)
 class MonthlyCap:
@@ -114,20 +81,11 @@ class MonthlyCap:
     is_default_ceiling: bool = False
 
 
-@dataclass(frozen=True)
-class AutoReloadCard:
-    kind: str  # "canonical" | "distinct" | "none"
-    payment_method_id: Optional[str] = None
-    brand: Optional[str] = None
-    last4: Optional[str] = None
-
-
 @dataclass(frozen=True)
 class AutoReload:
     enabled: bool = False
     threshold_usd: Optional[Decimal] = None
     reload_to_usd: Optional[Decimal] = None
-    card: Optional[AutoReloadCard] = None
 
 
 @dataclass(frozen=True)
@@ -142,8 +100,7 @@ class BillingState:
     org_id: Optional[str] = None
     org_slug: Optional[str] = None
     org_name: Optional[str] = None
-    role: Optional[str] = None  # "OWNER" | "ADMIN" | "FINANCE_ADMIN" | "SECURITY_ADMIN" | "MEMBER"
-    can_change_plan_raw: Optional[bool] = None
+    role: Optional[str] = None  # "OWNER" | "ADMIN" | "MEMBER"
     balance_usd: Optional[Decimal] = None
     cli_billing_enabled: bool = False
     charge_presets: tuple[Decimal, ...] = ()
@@ -158,33 +115,17 @@ class BillingState:
 
     @property
     def is_admin(self) -> bool:
-        """Deprecated/display only — a legacy OWNER/ADMIN check.
-
-        NOT a capability check; use :attr:`can_change_plan` for gating billing
-        plan-change actions.
-        """
+        """True for OWNER/ADMIN — the roles that can manage billing."""
         return (self.role or "").upper() in ("OWNER", "ADMIN")
 
-    @property
-    def can_change_plan(self) -> bool:
-        """Server capability when supplied; otherwise the legacy role fallback."""
-        if self.can_change_plan_raw is not None:
-            return self.can_change_plan_raw
-        return self.is_admin
-
     @property
     def can_charge(self) -> bool:
         """True when the UI should offer charge/auto-reload actions.
 
-        Uses the server-granted plan-change capability (``can_change_plan``,
-        which itself falls back to the legacy OWNER/ADMIN role check when the
-        server omits ``canChangePlan``) AND the per-org kill-switch. This lets
-        the server grant charge capability to non-OWNER/ADMIN roles (e.g.
-        FINANCE_ADMIN) via ``canChangePlan``, instead of hard-coding the
-        deprecated 3-role admin check. (The server still enforces; this is
-        just for graying out actions the user can't take.)
+        Admin role AND the per-org kill-switch on. (The server still enforces;
+        this is just for graying out actions the user can't take.)
         """
-        return self.can_change_plan and self.cli_billing_enabled
+        return self.is_admin and self.cli_billing_enabled
 
 
 def _parse_card(raw: Any) -> Optional[CardInfo]:
@@ -192,13 +133,9 @@ def _parse_card(raw: Any) -> Optional[CardInfo]:
         return None
     brand = raw.get("brand")
     last4 = raw.get("last4")
-    if not (isinstance(brand, str) and isinstance(last4, str)):
-        return None
-    # Post-resolver fields — all optional so both payload generations parse.
-    resolved_via = raw.get("resolvedVia")
-    if not isinstance(resolved_via, str):
-        resolved_via = None
-    return CardInfo(brand=brand, last4=last4, resolved_via=resolved_via)
+    if isinstance(brand, str) and isinstance(last4, str):
+        return CardInfo(brand=brand, last4=last4)
+    return None
 
 
 def _parse_monthly_cap(raw: Any) -> Optional[MonthlyCap]:
@@ -218,27 +155,6 @@ def _parse_auto_reload(raw: Any) -> Optional[AutoReload]:
         enabled=bool(raw.get("enabled")),
         threshold_usd=parse_money(raw.get("thresholdUsd")),
         reload_to_usd=parse_money(raw.get("reloadToUsd")),
-        card=_parse_auto_reload_card(raw.get("card")),
-    )
-
-
-def _parse_auto_reload_card(raw: Any) -> Optional[AutoReloadCard]:
-    if not isinstance(raw, dict):
-        return None
-    kind = raw.get("kind")
-    if kind not in ("canonical", "distinct", "none"):
-        return None
-    if kind in ("canonical", "none"):
-        return AutoReloadCard(kind=kind)
-
-    payment_method_id = raw.get("paymentMethodId")
-    brand = raw.get("brand")
-    last4 = raw.get("last4")
-    return AutoReloadCard(
-        kind=kind,
-        payment_method_id=payment_method_id if isinstance(payment_method_id, str) else None,
-        brand=brand if isinstance(brand, str) else None,
-        last4=last4 if isinstance(last4, str) else None,
     )
 
 
@@ -263,11 +179,6 @@ def billing_state_from_payload(
         org_slug=org.get("slug"),
         org_name=org.get("name"),
         role=org.get("role"),
-        can_change_plan_raw=(
-            payload.get("canChangePlan")
-            if isinstance(payload.get("canChangePlan"), bool)
-            else None
-        ),
         balance_usd=parse_money(payload.get("balanceUsd")),
         cli_billing_enabled=bool(payload.get("cliBillingEnabled")),
         charge_presets=tuple(presets),
@@ -291,15 +202,7 @@ def build_billing_state(*, timeout: float = 15.0) -> BillingState:
     Returns ``BillingState(logged_in=False)`` when not logged in. On a portal/HTTP
     failure, returns ``logged_in=False`` with ``error`` set so the surface can show
     a clear message rather than crashing.
-
-    Dev override: ``HERMES_DEV_BILLING_FIXTURE`` short-circuits to a fixture so the
-    card-on-file / admin / scope states are testable offline (mirrors
-    ``HERMES_DEV_CREDITS_FIXTURE`` for the usage model).
     """
-    fixture = _dev_fixture_billing_state()
-    if fixture is not None:
-        return fixture
-
     try:
         from hermes_cli.nous_billing import (
             BillingAuthError,
@@ -340,72 +243,6 @@ def _fallback_portal_url(base: str) -> str:
     return f"{base.rstrip('/')}/billing?topup=open"
 
 
-# =============================================================================
-# Dev fixtures (throwaway scaffolding — env-var driven, no live portal)
-# =============================================================================
-
-
-def _dev_fixture_billing_state() -> Optional[BillingState]:
-    """Map ``HERMES_DEV_BILLING_FIXTURE`` to a :class:`BillingState` for offline UX.
-
-    Recognized names::
-
-        nocard           logged in · billing on · admin · NO card on file
-        card             card on file · auto-reload off
-        card-autoreload  card on file · auto-reload on
-        notadmin         logged in · MEMBER role (billing actions disabled)
-        billing-off      logged in · admin · per-org kill-switch OFF
-        logged-out       not logged in
-
-    Returns ``None`` when the env var is unset (the real portal path runs).
-    Mirrors ``HERMES_DEV_CREDITS_FIXTURE``; the usage *bar* still comes from
-    ``HERMES_DEV_CREDITS_FIXTURE`` (set both to pair a bar with a billing state).
-    """
-    name = (os.getenv("HERMES_DEV_BILLING_FIXTURE") or "").strip().lower()
-    if not name:
-        return None
-
-    # Shared fixture portal host (matches subscription_view._DEV_FIXTURE_PORTAL —
-    # prod host, not staging; the ?topup=open suffix is the /topup deep-link).
-    portal = "https://portal.nousresearch.com/billing?topup=open"
-    common: dict[str, Any] = dict(
-        org_id="org_acme",
-        org_slug="acme",
-        org_name="Acme Inc",
-        role="OWNER",
-        balance_usd=Decimal("3.40"),
-        cli_billing_enabled=True,
-        charge_presets=(Decimal("10"), Decimal("25"), Decimal("50")),
-        min_usd=Decimal("5"),
-        max_usd=Decimal("500"),
-        portal_url=portal,
-    )
-    card = CardInfo(brand="Visa", last4="4242")
-    autoreload_on = AutoReload(enabled=True, threshold_usd=Decimal("5"), reload_to_usd=Decimal("25"))
-
-    if name in ("logged-out", "logged_out", "loggedout"):
-        return BillingState(logged_in=False)
-    if name == "nocard":
-        return BillingState(logged_in=True, card=None, **common)
-    if name == "card":
-        return BillingState(logged_in=True, card=card, **common)
-    if name in ("card-sub", "card_sub"):
-        # Post-resolver: the card came from the subscription (provenance label).
-        _sub_card = CardInfo(brand="Visa", last4="4242", resolved_via="subPin")
-        return BillingState(logged_in=True, card=_sub_card, **common)
-    if name in ("card-autoreload", "card_autoreload", "autoreload"):
-        return BillingState(logged_in=True, card=card, auto_reload=autoreload_on, **common)
-    if name in ("notadmin", "not-admin", "member"):
-        opts = {**common, "role": "MEMBER"}
-        return BillingState(logged_in=True, card=card, **opts)
-    if name in ("billing-off", "billing_off", "off"):
-        opts = {**common, "cli_billing_enabled": False}
-        return BillingState(logged_in=True, card=None, **opts)
-
-    # Unknown name → logged-out so the misconfiguration is visible.
-    return BillingState(logged_in=False, error=f"unknown HERMES_DEV_BILLING_FIXTURE: {name}")
-
-
 # =============================================================================
 # Idempotency
 # =============================================================================

agent/bounded_response.py

@@ -1,148 +0,0 @@
-"""Bounded reads of HTTP error response bodies.
-
-When a provider returns a non-OK status on a *streaming* request, Hermes reads
-the response body to build a useful diagnostic error. A bare ``response.read()``
-on a streaming httpx response is unbounded in two dangerous ways:
-
-1. A server can declare (or stream) an arbitrarily large body, so the read can
-   balloon memory.
-2. A server can open the body and then stall forever (no ``Content-Length``,
-   no further bytes), so the read hangs the agent indefinitely.
-
-Both are realistic against a misbehaving proxy, a hijacked endpoint, or a
-provider having a bad day. The diagnostic body is only ever shown to the user
-truncated to a few hundred characters, so reading megabytes — or blocking
-forever — buys nothing.
-
-``read_streaming_error_body`` bounds the read to a byte cap and enforces a
-hard wall-clock deadline, returning the decoded text snippet. Callers pass the
-returned text into their existing error builders instead of touching
-``response.text`` (which would be unbounded / would raise after a partial
-stream read).
-
-A subtlety the implementation must respect: ``httpx``'s ``iter_bytes()`` blocks
-*inside* the C/socket read while waiting for the next chunk. A wall-clock check
-placed only between yielded chunks cannot interrupt a server that opens the
-body and then stalls mid-chunk — control never returns to Python until httpx's
-own (often 30s+) read timeout fires. To guarantee a bounded stop regardless of
-socket behavior, the read runs on a daemon worker thread and the caller waits
-on it with a hard deadline; on timeout we close the response (which unblocks /
-cancels the read) and return whatever partial bytes were collected.
-
-Ported and adapted from openclaw/openclaw#95108 ("bound Anthropic error
-streams"), generalized to cover Hermes's three streaming error-body sites
-(native Gemini, Gemini Cloud Code, Antigravity Cloud Code).
-"""
-
-from __future__ import annotations
-
-import logging
-import threading
-from typing import List, Optional
-
-import httpx
-
-logger = logging.getLogger(__name__)
-
-# Defaults chosen to comfortably hold any real provider error envelope (Google
-# RPC error JSON, Anthropic error JSON) while rejecting pathological bodies.
-DEFAULT_ERROR_BODY_MAX_BYTES = 64 * 1024
-# Hard wall-clock deadline for the whole bounded read. A streaming error body
-# that does not finish within this window is abandoned and the connection is
-# closed; we keep whatever partial bytes arrived.
-DEFAULT_ERROR_BODY_TIMEOUT_S = 10.0
-
-
-def read_streaming_error_body(
-    response: httpx.Response,
-    *,
-    max_bytes: int = DEFAULT_ERROR_BODY_MAX_BYTES,
-    timeout_s: float = DEFAULT_ERROR_BODY_TIMEOUT_S,
-) -> str:
-    """Read a non-OK streaming response body with a byte cap and a hard deadline.
-
-    Returns the decoded body text (UTF-8, errors replaced), truncated to
-    ``max_bytes``. Never raises: any transport error, stall, or oversize
-    condition is swallowed and the best-effort partial text (or an empty
-    string) is returned, because this runs on the error path and must not
-    mask the original HTTP failure with a read error.
-
-    The byte cap protects against huge bodies; the wall-clock deadline (enforced
-    via a worker thread so it can interrupt a socket read that stalls mid-chunk)
-    protects against bodies that open and then hang.
-    """
-    chunks: List[bytes] = []
-    state = {"truncated": False}
-    done = threading.Event()
-
-    def _drain() -> None:
-        total = 0
-        try:
-            for chunk in response.iter_bytes():
-                if not chunk:
-                    continue
-                remaining = max_bytes - total
-                if remaining <= 0:
-                    state["truncated"] = True
-                    break
-                if len(chunk) > remaining:
-                    chunks.append(chunk[:remaining])
-                    total += remaining
-                    state["truncated"] = True
-                    break
-                chunks.append(chunk)
-                total += len(chunk)
-        except Exception as exc:  # noqa: BLE001 - error path must not raise
-            logger.debug("bounded error-body read failed: %s", exc)
-        finally:
-            done.set()
-
-    worker = threading.Thread(
-        target=_drain, name="bounded-error-body-read", daemon=True
-    )
-    worker.start()
-    finished = done.wait(timeout=timeout_s)
-
-    if not finished:
-        logger.debug(
-            "bounded error-body read: hard timeout after %.1fs (%d bytes so far)",
-            timeout_s,
-            sum(len(c) for c in chunks),
-        )
-        # Closing the response cancels the in-flight socket read, letting the
-        # worker thread unwind. We do not join (it is a daemon and may be
-        # blocked in C); the partial `chunks` collected so far are returned.
-        _safe_close(response)
-    else:
-        _safe_close(response)
-
-    if state["truncated"]:
-        logger.debug(
-            "bounded error-body read: capped at %d bytes (max=%d)",
-            sum(len(c) for c in chunks),
-            max_bytes,
-        )
-    return b"".join(chunks).decode("utf-8", errors="replace")
-
-
-def _safe_close(response: httpx.Response) -> None:
-    try:
-        response.close()
-    except Exception:  # noqa: BLE001
-        pass
-
-
-def read_error_body_or_default(
-    response: httpx.Response,
-    *,
-    max_bytes: int = DEFAULT_ERROR_BODY_MAX_BYTES,
-    timeout_s: float = DEFAULT_ERROR_BODY_TIMEOUT_S,
-) -> Optional[str]:
-    """Like ``read_streaming_error_body`` but returns ``None`` on empty body.
-
-    Convenience for callers that distinguish "no body" from "empty string".
-    """
-    text = read_streaming_error_body(
-        response, max_bytes=max_bytes, timeout_s=timeout_s
-    )
-    return text or None

agent/codex_responses_adapter.py

@@ -288,13 +288,6 @@ _RESPONSES_BUILTIN_TOOL_TYPES = {
 
 _RESPONSE_MESSAGE_STATUSES = {"completed", "incomplete", "in_progress"}
 
-# The Responses API rejects input[].id longer than this with a non-retryable
-# HTTP 400 ("string too long"). Codex-issued assistant message ids are
-# server-assigned base64 blobs that can run 400+ chars, while Hermes-minted
-# ids (msg_...) stay well under this cap and are worth keeping for
-# prefix-cache hits. Drop only the oversized ones on replay.
-_MAX_RESPONSES_ITEM_ID_LENGTH = 64
-
 
 def _normalize_responses_message_status(value: Any, *, default: str = "completed") -> str:
     """Normalize a Responses assistant message status for replay.
@@ -314,7 +307,6 @@ def _chat_messages_to_responses_input(
     messages: List[Dict[str, Any]],
     *,
     is_xai_responses: bool = False,
-    is_github_responses: bool = False,
     replay_encrypted_reasoning: bool = True,
     current_issuer_kind: Optional[str] = None,
 ) -> List[Dict[str, Any]]:
@@ -339,16 +331,6 @@ def _chat_messages_to_responses_input(
     items from the conversation history and threads ``replay_enabled=False``
     through this converter so subsequent turns send no reasoning items.
 
-    ``is_github_responses`` drops the ``id`` field from replayed
-    ``codex_message_items`` regardless of length. The Copilot backend
-    (api.githubcopilot.com/responses) binds these ids to a specific
-    backend "connection" — credential-pool rotation, a gateway restart,
-    or routine load-balancer churn between turns all invalidate it — and
-    rejects a stale id with HTTP 401 "input item ID does not belong to
-    this connection" even for short ids (see #32716). ``phase``/
-    ``status``/``content`` are still replayed; only ``id`` is unsafe to
-    reuse across a Copilot connection.
-
     ``current_issuer_kind`` enables a per-item cross-issuer guard. The
     Responses API's ``encrypted_content`` blob is decryptable only by the
     endpoint that minted it — replaying a Codex-issued blob against xAI
@@ -481,14 +463,8 @@ def _chat_messages_to_responses_input(
                             "content": normalized_content_parts,
                         }
                         item_id = raw_item.get("id")
-                        if (
-                            not is_github_responses
-                            and isinstance(item_id, str)
-                            and item_id.strip()
-                        ):
-                            stripped_id = item_id.strip()
-                            if len(stripped_id) <= _MAX_RESPONSES_ITEM_ID_LENGTH:
-                                replay_item["id"] = stripped_id
+                        if isinstance(item_id, str) and item_id.strip():
+                            replay_item["id"] = item_id.strip()
                         phase = raw_item.get("phase")
                         if isinstance(phase, str) and phase.strip():
                             replay_item["phase"] = phase.strip()
@@ -600,11 +576,7 @@ def _chat_messages_to_responses_input(
 # Input preflight / validation
 # ---------------------------------------------------------------------------
 
-def _preflight_codex_input_items(
-    raw_items: Any,
-    *,
-    is_github_responses: bool = False,
-) -> List[Dict[str, Any]]:
+def _preflight_codex_input_items(raw_items: Any) -> List[Dict[str, Any]]:
     if not isinstance(raw_items, list):
         raise ValueError("Codex Responses input must be a list of input items.")
 
@@ -745,14 +717,8 @@ def _preflight_codex_input_items(
                 "content": normalized_content,
             }
             item_id = item.get("id")
-            if (
-                not is_github_responses
-                and isinstance(item_id, str)
-                and item_id.strip()
-            ):
-                stripped_id = item_id.strip()
-                if len(stripped_id) <= _MAX_RESPONSES_ITEM_ID_LENGTH:
-                    normalized_item["id"] = stripped_id
+            if isinstance(item_id, str) and item_id.strip():
+                normalized_item["id"] = item_id.strip()
             phase = item.get("phase")
             if isinstance(phase, str) and phase.strip():
                 normalized_item["phase"] = phase.strip()
@@ -824,7 +790,6 @@ def _preflight_codex_api_kwargs(
     api_kwargs: Any,
     *,
     allow_stream: bool = False,
-    is_github_responses: bool = False,
 ) -> Dict[str, Any]:
     if not isinstance(api_kwargs, dict):
         raise ValueError("Codex Responses request must be a dict.")
@@ -846,10 +811,7 @@ def _preflight_codex_api_kwargs(
         instructions = str(instructions)
     instructions = instructions.strip() or DEFAULT_AGENT_IDENTITY
 
-    normalized_input = _preflight_codex_input_items(
-        api_kwargs.get("input"),
-        is_github_responses=is_github_responses,
-    )
+    normalized_input = _preflight_codex_input_items(api_kwargs.get("input"))
 
     tools = api_kwargs.get("tools")
     normalized_tools = None
@@ -912,8 +874,7 @@ def _preflight_codex_api_kwargs(
     allowed_keys = {
         "model", "instructions", "input", "tools", "store",
         "reasoning", "include", "max_output_tokens", "temperature",
-        "tool_choice", "parallel_tool_calls", "prompt_cache_key",
-        "prompt_cache_retention", "service_tier",
+        "tool_choice", "parallel_tool_calls", "prompt_cache_key", "service_tier",
         "extra_headers", "extra_body", "timeout",
     }
     normalized: Dict[str, Any] = {
@@ -951,13 +912,8 @@ def _preflight_codex_api_kwargs(
     if isinstance(temperature, (int, float)):
         normalized["temperature"] = float(temperature)
 
-    # Pass through cache routing/retention and tool-dispatch hints.
-    for passthrough_key in (
-        "tool_choice",
-        "parallel_tool_calls",
-        "prompt_cache_key",
-        "prompt_cache_retention",
-    ):
+    # Pass through tool_choice, parallel_tool_calls, prompt_cache_key
+    for passthrough_key in ("tool_choice", "parallel_tool_calls", "prompt_cache_key"):
         val = api_kwargs.get(passthrough_key)
         if val is not None:
             normalized[passthrough_key] = val
@@ -1124,22 +1080,6 @@ def _normalize_codex_response(
     differs from the one that minted the encrypted_content blob and drop
     the item instead of triggering HTTP 400 invalid_encrypted_content.
     """
-    response_status = getattr(response, "status", None)
-    if isinstance(response_status, str):
-        response_status = response_status.strip().lower()
-    else:
-        response_status = None
-
-    incomplete_details = getattr(response, "incomplete_details", None)
-    incomplete_reason = ""
-    if isinstance(incomplete_details, dict):
-        incomplete_reason = str(incomplete_details.get("reason") or "").strip().lower()
-    elif incomplete_details is not None:
-        incomplete_reason = str(getattr(incomplete_details, "reason", "") or "").strip().lower()
-    response_incomplete_content_filter = (
-        response_status == "incomplete" and incomplete_reason == "content_filter"
-    )
-
     output = getattr(response, "output", None)
     if not isinstance(output, list) or not output:
         # The Codex backend can return empty output when the answer was
@@ -1156,18 +1096,15 @@ def _normalize_codex_response(
                 content=[SimpleNamespace(type="output_text", text=out_text.strip())],
             )]
             response.output = output
-        elif response_incomplete_content_filter:
-            # This is a deterministic provider safety block, not a partial
-            # answer. Synthesize an empty message so finish_reason below becomes
-            # content_filter and the conversation loop can fallback/surface it
-            # instead of burning three continuation attempts.
-            output = [SimpleNamespace(
-                type="message", role="assistant", status="completed", content=[]
-            )]
-            response.output = output
         else:
             raise RuntimeError("Responses API returned no output items")
 
+    response_status = getattr(response, "status", None)
+    if isinstance(response_status, str):
+        response_status = response_status.strip().lower()
+    else:
+        response_status = None
+
     if response_status in {"failed", "cancelled"}:
         error_obj = getattr(response, "error", None)
         error_msg = _format_responses_error(error_obj, response_status)
@@ -1229,28 +1166,15 @@ def _normalize_codex_response(
         if item_type == "message":
             item_phase = getattr(item, "phase", None)
             normalized_phase = None
-            is_commentary_phase = False
             if isinstance(item_phase, str):
                 normalized_phase = item_phase.strip().lower()
                 if normalized_phase in {"commentary", "analysis"}:
                     saw_commentary_phase = True
-                    is_commentary_phase = True
                 elif normalized_phase in {"final_answer", "final"}:
                     saw_final_answer_phase = True
             message_text = _extract_responses_message_text(item)
             if message_text:
-                # Responses ``commentary``/``analysis`` phase text is mid-turn
-                # preamble/progress narration, never the turn's final answer
-                # (Codex CLI excludes it from last-message extraction; issues
-                # #24933 / #41293).  Keep it out of assistant content so it
-                # can't be concatenated into — or leak as — the final response,
-                # but surface it through the reasoning channel so the CLI/
-                # gateway display it like thinking text.  The exact message
-                # item is still preserved below for replay/cache continuity.
-                if is_commentary_phase:
-                    reasoning_parts.append(message_text)
-                else:
-                    content_parts.append(message_text)
+                content_parts.append(message_text)
                 raw_message_item: Dict[str, Any] = {
                     "type": "message",
                     "role": "assistant",
@@ -1345,11 +1269,7 @@ def _normalize_codex_response(
             ))
 
     final_text = "\n".join([p for p in content_parts if p]).strip()
-    if (
-        not final_text
-        and hasattr(response, "output_text")
-        and not (saw_commentary_phase and not saw_final_answer_phase)
-    ):
+    if not final_text and hasattr(response, "output_text"):
         out_text = getattr(response, "output_text", "")
         if isinstance(out_text, str):
             final_text = out_text.strip()
@@ -1385,45 +1305,6 @@ def _normalize_codex_response(
         # so the model keeps its chain-of-thought on the retry.
         final_text = ""
 
-    # ── Reasoning-channel answer salvage (xAI grok) ──────────────
-    # grok-4.x on the xAI /v1/responses surface sometimes emits its final
-    # answer inside the reasoning item instead of as a ``message`` output
-    # item, marking where the answer starts with grok's internal
-    # ``<response>`` delimiter.  Without salvage, the reasoning-only rule
-    # below classifies the turn ``incomplete`` — and because reasoning
-    # items on this surface carry no ``encrypted_content``, the interim
-    # message replays as nothing, so every continuation request is
-    # byte-identical to the one that just failed.  The turn burns its 3
-    # retries and dies with "Codex response remained incomplete after 3
-    # continuation attempts" even though the answer was produced on the
-    # first attempt.  Observed live with grok-4.20 on xai-oauth
-    # (2026-07-13).  Promote the delimited tail to assistant content and
-    # keep the untagged prefix as thinking text.
-    if (
-        issuer_kind == "xai_responses"
-        and not final_text
-        and not tool_calls
-        and reasoning_parts
-    ):
-        joined_reasoning = "\n\n".join(reasoning_parts)
-        marker = joined_reasoning.rfind("<response>")
-        if marker != -1:
-            salvaged = joined_reasoning[marker + len("<response>"):]
-            closing = salvaged.find("</response>")
-            if closing != -1:
-                salvaged = salvaged[:closing]
-            salvaged = salvaged.strip()
-            if salvaged:
-                logger.warning(
-                    "xAI response delivered its final answer inside the "
-                    "reasoning channel (<response> delimiter); promoting "
-                    "%d chars to assistant content.",
-                    len(salvaged),
-                )
-                final_text = salvaged
-                reasoning_prefix = joined_reasoning[:marker].strip()
-                reasoning_parts = [reasoning_prefix] if reasoning_prefix else []
-
     assistant_message = SimpleNamespace(
         content=final_text,
         tool_calls=tool_calls,
@@ -1436,8 +1317,6 @@ def _normalize_codex_response(
 
     if tool_calls:
         finish_reason = "tool_calls"
-    elif response_incomplete_content_filter:
-        finish_reason = "content_filter"
     elif leaked_tool_call_text:
         finish_reason = "incomplete"
     elif saw_streaming_or_item_incomplete:
@@ -1446,28 +1325,12 @@ def _normalize_codex_response(
         finish_reason = "incomplete"
     elif (reasoning_items_raw or reasoning_parts or saw_reasoning_item) and not final_text:
         # Response contains only reasoning (encrypted thinking state and/or
-        # human-readable summary) with no visible content or tool calls.
-        #
-        # For the specially-handled backends (Codex, xAI, GitHub/Copilot),
-        # reasoning-only with status="completed" means "the model is still
-        # thinking and needs another turn" — treat it as incomplete so the
-        # Codex continuation path retries instead of falling into the
-        # empty-content retry loop.
-        #
-        # For all other backends (other:<base_url>, etc.), trust the provider's
-        # own response.status signal. When status == "completed" and no items
-        # are queued/in_progress/incomplete, reasoning alone is a valid final
-        # state — forcing "incomplete" causes multi-minute stalls as the
-        # continuation path re-issues calls (3 retries × up to 240s each).
-        # See https://github.com/NousResearch/hermes-agent/issues/64434
-        if response_status == "completed" and issuer_kind not in (
-            "codex_backend",
-            "xai_responses",
-            "github_responses",
-        ):
-            finish_reason = "stop"
-        else:
-            finish_reason = "incomplete"
+        # human-readable summary) with no visible content or tool calls. The
+        # model is still thinking and needs another turn to produce the actual
+        # answer. Marking this as "stop" would send it into the empty-content
+        # retry loop which burns retries then fails — treat it as incomplete so
+        # the Codex continuation path handles it correctly.
+        finish_reason = "incomplete"
     else:
         finish_reason = "stop"
     return assistant_message, finish_reason

agent/codex_runtime.py

@@ -16,18 +16,70 @@ compatibility.
 
 from __future__ import annotations
 
-import json
 import logging
 import os
 import time
 from types import SimpleNamespace
-from typing import Any, Callable, Dict, List
-
-from agent.stream_single_writer import claim_stream_writer, stream_writer_is_current
+from typing import Any, Dict, List
 
 logger = logging.getLogger(__name__)
 
 
+def _codex_note_to_tool_progress(note: dict) -> tuple[str, str, dict] | None:
+    """Map a Codex app-server ``item/started`` notification to a Hermes
+    tool-progress event ``(tool_name, preview, args)``.
+
+    The Codex app-server runtime processes ``item/started`` notifications for
+    command execution, file changes, and MCP/dynamic tool calls, but never
+    surfaced them as Hermes tool-progress events — so gateways (Telegram, etc.)
+    showed no verbose "running X" breadcrumbs on this route while every other
+    provider did (#38835). Returns None for items that aren't tool-shaped.
+    """
+    if not isinstance(note, dict) or note.get("method") != "item/started":
+        return None
+    params = note.get("params") or {}
+    item = params.get("item") or {}
+    if not isinstance(item, dict):
+        return None
+
+    item_type = item.get("type") or ""
+    if item_type == "commandExecution":
+        command = item.get("command") or ""
+        return "exec_command", command, {"command": command, "cwd": item.get("cwd") or ""}
+
+    if item_type == "fileChange":
+        changes = item.get("changes") or []
+        preview = "file changes"
+        if isinstance(changes, list) and changes:
+            paths = [
+                str(change.get("path"))
+                for change in changes
+                if isinstance(change, dict) and change.get("path")
+            ]
+            if paths:
+                preview = ", ".join(paths[:3])
+                if len(paths) > 3:
+                    preview += f", +{len(paths) - 3} more"
+        return "apply_patch", preview, {"changes": changes}
+
+    if item_type == "mcpToolCall":
+        server = item.get("server") or "mcp"
+        tool = item.get("tool") or "unknown"
+        args = item.get("arguments") or {}
+        if not isinstance(args, dict):
+            args = {"arguments": args}
+        return f"mcp.{server}.{tool}", tool, args
+
+    if item_type == "dynamicToolCall":
+        tool = item.get("tool") or "unknown"
+        args = item.get("arguments") or {}
+        if not isinstance(args, dict):
+            args = {"arguments": args}
+        return tool, tool, args
+
+    return None
+
+
 def _coerce_usage_int(value: Any) -> int:
     if isinstance(value, bool):
         return 0
@@ -61,15 +113,6 @@ def _record_codex_app_server_usage(agent, turn) -> dict[str, Any]:
 
     usage = getattr(turn, "token_usage_last", None)
     if not isinstance(usage, dict) or not usage:
-        compressor = getattr(agent, "context_compressor", None)
-        if (
-            compressor is not None
-            and getattr(compressor, "awaiting_real_usage_after_compression", False)
-        ):
-            # No usage means this turn cannot adjudicate the pending compaction.
-            # Consume the marker so a later unrelated reading is not charged to
-            # it and preflight deferral cannot stay latched indefinitely.
-            compressor.update_from_response({})
         if agent._session_db and agent.session_id:
             try:
                 if not agent._session_db_created:
@@ -77,9 +120,6 @@ def _record_codex_app_server_usage(agent, turn) -> dict[str, Any]:
                 agent._session_db.update_token_counts(
                     agent.session_id,
                     model=agent.model,
-                    billing_provider=agent.provider,
-                    billing_base_url=agent.base_url,
-                    billing_mode="subscription_included",
                     api_call_count=1,
                 )
             except Exception as exc:
@@ -188,430 +228,6 @@ def _record_codex_app_server_usage(agent, turn) -> dict[str, Any]:
     }
 
 
-def _record_codex_app_server_compaction(
-    agent,
-    turn,
-    *,
-    approx_tokens: int | None = None,
-    force: bool = False,
-) -> bool:
-    """Record a Codex-native context compaction boundary in Hermes state.
-
-    The app-server owns the compacted thread context, so Hermes should not
-    rewrite local transcript rows here; state.db records the boundary via the
-    session event/usage counters while preserving the visible transcript.
-    """
-    if not force and not getattr(turn, "compacted", False):
-        return False
-
-    thread_id = getattr(turn, "thread_id", None) or ""
-    turn_id = getattr(turn, "turn_id", None) or ""
-    logger.info(
-        "codex app-server compaction observed: session=%s thread=%s turn=%s force=%s",
-        getattr(agent, "session_id", None) or "none",
-        thread_id,
-        turn_id,
-        force,
-    )
-    if not force:
-        try:
-            from agent.conversation_compression import COMPACTION_STATUS
-
-            agent._emit_status(COMPACTION_STATUS)
-        except Exception:
-            pass
-
-    compressor = getattr(agent, "context_compressor", None)
-    if compressor is not None:
-        compressor.compression_count = getattr(
-            compressor, "compression_count", 0
-        ) + 1
-        compressor.last_compression_rough_tokens = approx_tokens or 0
-        # The app server has already completed a real compaction boundary. Its
-        # usage update (when supplied) is therefore the same real-vs-real
-        # effectiveness verdict used by the normal compression path.
-        record_boundary = getattr(
-            type(compressor), "record_completed_compaction", None
-        )
-        if callable(record_boundary):
-            # Codex owns this summary. A prior Hermes deterministic-fallback
-            # flag must not leak into the native boundary's quality verdict.
-            record_boundary(compressor, used_fallback=False)
-        elif hasattr(compressor, "_verify_compaction_cleared_threshold"):
-            compressor._verify_compaction_cleared_threshold = True
-        if not getattr(turn, "token_usage_last", None):
-            compressor.last_prompt_tokens = -1
-            compressor.last_completion_tokens = 0
-            compressor.awaiting_real_usage_after_compression = True
-
-    agent._last_compaction_in_place = False
-    try:
-        if getattr(agent, "event_callback", None):
-            agent.event_callback(
-                "session:compress",
-                {
-                    "platform": getattr(agent, "platform", None) or "",
-                    "session_id": getattr(agent, "session_id", None) or "",
-                    "old_session_id": "",
-                    "in_place": False,
-                    "compression_count": getattr(
-                        compressor, "compression_count", 0
-                    )
-                    if compressor is not None
-                    else 0,
-                    "runtime": "codex_app_server",
-                    "thread_id": thread_id,
-                    "turn_id": turn_id,
-                },
-            )
-    except Exception:
-        logger.debug("event_callback error on codex session:compress", exc_info=True)
-
-    return True
-
-
-# ---------------------------------------------------------------------------
-# Codex app-server → Hermes UI bridge (#33200)
-#
-# The codex_app_server runtime hands the entire turn to a subprocess and
-# bypasses the normal Hermes tool loop. Without this bridge gateway
-# adapters (Discord, Telegram, TUI) never see live tool-progress bubbles
-# or interim assistant commentary while codex is working — the user just
-# stares at a quiet channel until the final answer lands. The bridge
-# translates raw codex JSON-RPC notifications into the same three agent
-# callbacks the standard runtime fires:
-#   - tool_progress_callback("tool.started"|"tool.completed", name, ...)
-#   - _fire_stream_delta(text) for streaming agentMessage chunks
-#   - _emit_interim_assistant_message({...}) for completed agentMessages
-# ---------------------------------------------------------------------------
-
-# Codex item types that map to a Hermes tool_call in the projector (and
-# therefore deserve a tool_progress bubble pair). The projector lives in
-# agent/transports/codex_event_projector.py — keep these in sync so the
-# tool name shown in the UI matches the name recorded in messages.
-# webSearch is codex's built-in web search tool — it has no projector
-# entry (codex handles it internally) but still deserves a bubble.
-_CODEX_TOOL_ITEM_TYPES = frozenset(
-    {"commandExecution", "fileChange", "mcpToolCall", "dynamicToolCall", "webSearch"}
-)
-
-# Internal MCP server that wraps Hermes' native tools for codex. When
-# codex calls back through it, the inner dispatch runs in a SEPARATE
-# hermes-tools-mcp-server subprocess that has no access to the parent
-# agent's tool_progress_callback — so the inner call can never surface
-# its own native progress event. The codex-level mcpToolCall event IS
-# the display event for those calls; we strip the mcp.hermes-tools.*
-# namespacing and emit the bare tool name (web_search, browser_navigate,
-# vision_analyze, ...) since the user thinks of these as Hermes tools,
-# not as MCP calls.
-_INTERNAL_MCP_SERVER = "hermes-tools"
-
-
-def _codex_item_to_tool_name(item: dict) -> str:
-    """Synthetic Hermes tool name for a codex item. Mirrors
-    CodexEventProjector so the progress bubble and the projected
-    tool_calls entry use the same identifier."""
-    item_type = item.get("type") or ""
-    if item_type == "commandExecution":
-        return "exec_command"
-    if item_type == "fileChange":
-        return "apply_patch"
-    if item_type == "mcpToolCall":
-        server = item.get("server") or "mcp"
-        tool = item.get("tool") or "unknown"
-        if server == _INTERNAL_MCP_SERVER:
-            return tool
-        return f"mcp.{server}.{tool}"
-    if item_type == "dynamicToolCall":
-        return item.get("tool") or "dynamic"
-    if item_type == "webSearch":
-        return "web_search"
-    return item_type or "unknown"
-
-
-def _codex_item_to_args(item: dict) -> dict:
-    """Args dict surfaced to tool_progress_callback("tool.started", ...).
-    Mirrors the projector's _project_command / _project_file_change /
-    _project_mcp_tool_call / _project_dynamic_tool_call shapes."""
-    item_type = item.get("type") or ""
-    if item_type == "commandExecution":
-        return {"command": item.get("command") or "",
-                "cwd": item.get("cwd") or ""}
-    if item_type == "fileChange":
-        return {"changes": [
-            {"kind": (c.get("kind") or {}).get("type") or "update",
-             "path": c.get("path") or ""}
-            for c in (item.get("changes") or []) if isinstance(c, dict)
-        ]}
-    if item_type in {"mcpToolCall", "dynamicToolCall"}:
-        args = item.get("arguments") or {}
-        return args if isinstance(args, dict) else {"arguments": args}
-    if item_type == "webSearch":
-        return {"query": item.get("query") or ""}
-    return {}
-
-
-def _codex_item_to_preview(item: dict) -> Any:
-    """Short human-readable preview for the tool.started bubble. Returns
-    None when no useful preview is available (Hermes' UI tolerates None)."""
-    item_type = item.get("type") or ""
-    if item_type == "commandExecution":
-        cmd = item.get("command") or ""
-        return cmd[:120] if cmd else None
-    if item_type == "fileChange":
-        paths = [c.get("path") for c in (item.get("changes") or [])
-                 if isinstance(c, dict) and c.get("path")]
-        if not paths:
-            return None
-        preview = ", ".join(paths[:3])
-        if len(paths) > 3:
-            preview += f", +{len(paths) - 3} more"
-        return preview
-    if item_type in {"mcpToolCall", "dynamicToolCall"}:
-        args = item.get("arguments") or {}
-        if not isinstance(args, dict) or not args:
-            return None
-        try:
-            return json.dumps(args, ensure_ascii=False)[:120]
-        except (TypeError, ValueError):
-            return None
-    if item_type == "webSearch":
-        query = item.get("query") or ""
-        return query[:120] if query else None
-    return None
-
-
-def _codex_item_completion_payload(item: dict) -> tuple[str, bool]:
-    """Return (result_text, is_error) for a completed codex tool item.
-    Mirrors the projector's tool-result content so the bubble shows the
-    same outcome string that ends up in the messages list."""
-    item_type = item.get("type") or ""
-    if item_type == "commandExecution":
-        out = item.get("aggregatedOutput") or ""
-        exit_code = item.get("exitCode")
-        is_error = bool(exit_code is not None and exit_code != 0)
-        if is_error:
-            out = f"[exit {exit_code}]\n{out}"
-        return out, is_error
-    if item_type == "fileChange":
-        status = item.get("status") or "unknown"
-        n = len(item.get("changes") or [])
-        return (
-            f"apply_patch status={status}, {n} change(s)",
-            status not in {"completed", "applied", "success"},
-        )
-    if item_type == "mcpToolCall":
-        error = item.get("error")
-        if error:
-            return (
-                f"[error] {json.dumps(error, ensure_ascii=False)[:1000]}",
-                True,
-            )
-        result = item.get("result")
-        return (
-            json.dumps(result, ensure_ascii=False)[:4000]
-            if result is not None else "",
-            False,
-        )
-    if item_type == "dynamicToolCall":
-        content_items = item.get("contentItems") or []
-        if isinstance(content_items, list) and content_items:
-            return (
-                json.dumps(content_items, ensure_ascii=False)[:4000],
-                not bool(item.get("success", True)),
-            )
-        success = item.get("success", True)
-        return f"success={success}", not bool(success)
-    return "", False
-
-
-def make_codex_app_server_event_bridge(agent) -> Callable[[dict], None]:
-    """Build an ``on_event`` callback that wires codex app-server JSON-RPC
-    notifications into Hermes' gateway UI callbacks.
-
-    Returns a single-argument callable suitable for
-    ``CodexAppServerSession(on_event=...)``.
-
-    Translation map:
-      * ``item/started`` for tool-shaped items → ``tool_progress_callback(
-        "tool.started", name, preview, args)``
-      * ``item/completed`` for tool-shaped items → ``tool_progress_callback(
-        "tool.completed", name, None, None, duration=..., is_error=...,
-        result=...)``
-      * ``item/agentMessage/delta`` → ``_fire_stream_delta(text)`` so chat
-        adapters can render the assistant's reply as it streams.
-      * ``item/reasoning/delta`` → ``_fire_reasoning_delta(text)``
-      * ``item/completed`` for ``agentMessage`` →
-        ``_emit_interim_assistant_message({"role": "assistant",
-        "content": text})``. The gateway's ``already_streamed`` check
-        dedupes against any text the stream-delta callback already
-        rendered for the same message.
-
-    All callback invocations are guarded — a buggy display callback must
-    not tear down the codex turn loop. Errors are logged at DEBUG so the
-    notification stream keeps flowing regardless.
-    """
-    # item_id -> (tool_name, args, started_wall_time). Populated on
-    # item/started and consumed on item/completed so duration is correct
-    # even when codex doesn't report durationMs.
-    started: dict[str, tuple[str, dict, float]] = {}
-
-    def _stable_call_id(item: dict, name: str) -> str:
-        """Deterministic tool_call id mirroring CodexEventProjector, so a
-        live TUI tool card correlates with the same tool call after the
-        session is resumed and history is projected."""
-        from agent.transports.codex_event_projector import _deterministic_call_id
-
-        item_id = item.get("id") or ""
-        item_type = item.get("type") or ""
-        if item_type == "commandExecution":
-            return _deterministic_call_id("exec", item_id)
-        if item_type == "fileChange":
-            return _deterministic_call_id("apply_patch", item_id)
-        if item_type == "mcpToolCall":
-            server = item.get("server") or "mcp"
-            tool = item.get("tool") or "unknown"
-            return _deterministic_call_id(f"mcp__{server}__{tool}", item_id)
-        if item_type == "dynamicToolCall":
-            tool = item.get("tool") or "unknown"
-            return _deterministic_call_id(f"dyn_{tool}", item_id)
-        return _deterministic_call_id(name, item_id)
-
-    def _fire_tool_started(item: dict) -> None:
-        item_id = item.get("id") or ""
-        name = _codex_item_to_tool_name(item)
-        args = _codex_item_to_args(item)
-        if item_id:
-            started[item_id] = (name, args, time.monotonic())
-        cb = getattr(agent, "tool_progress_callback", None)
-        if cb is not None:
-            try:
-                cb("tool.started", name, _codex_item_to_preview(item), args)
-            except Exception:
-                logger.debug(
-                    "tool_progress_callback raised on tool.started for %s",
-                    name, exc_info=True,
-                )
-        # Authoritative stable-ID tool card (TUI / desktop). Fires
-        # alongside tool_progress so surfaces that render structured tool
-        # cards (not just progress bubbles) stay correlated with the
-        # projected history entry after a resume.
-        start_cb = getattr(agent, "tool_start_callback", None)
-        if start_cb is not None:
-            try:
-                start_cb(_stable_call_id(item, name), name, args)
-            except Exception:
-                logger.debug(
-                    "tool_start_callback raised for %s", name, exc_info=True,
-                )
-
-    def _fire_tool_completed(item: dict) -> None:
-        item_id = item.get("id") or ""
-        name = _codex_item_to_tool_name(item)
-        prior = started.pop(item_id, None)
-        # Prefer codex's own durationMs when present so the bubble shows
-        # exact tool wall-time; fall back to our started timestamp; fall
-        # back to None if we never saw an item/started (some codex
-        # versions only emit completed for fast items).
-        duration: Any = None
-        codex_ms = item.get("durationMs")
-        if isinstance(codex_ms, (int, float)) and codex_ms >= 0:
-            duration = codex_ms / 1000.0
-        elif prior is not None:
-            duration = time.monotonic() - prior[2]
-        result, is_error = _codex_item_completion_payload(item)
-        cb = getattr(agent, "tool_progress_callback", None)
-        if cb is not None:
-            try:
-                cb("tool.completed", name, None, None,
-                   duration=duration, is_error=is_error, result=result)
-            except Exception:
-                logger.debug(
-                    "tool_progress_callback raised on tool.completed for %s",
-                    name, exc_info=True,
-                )
-        complete_cb = getattr(agent, "tool_complete_callback", None)
-        if complete_cb is not None:
-            args = prior[1] if prior is not None else _codex_item_to_args(item)
-            try:
-                complete_cb(_stable_call_id(item, name), name, args, result)
-            except Exception:
-                logger.debug(
-                    "tool_complete_callback raised for %s", name, exc_info=True,
-                )
-
-    def _fire_text_delta(params: dict) -> None:
-        text = params.get("delta") or params.get("text") or ""
-        if not isinstance(text, str) or not text:
-            return
-        fn = getattr(agent, "_fire_stream_delta", None)
-        if fn is None:
-            return
-        try:
-            fn(text)
-        except Exception:
-            logger.debug("_fire_stream_delta raised", exc_info=True)
-
-    def _fire_reasoning_delta(params: dict) -> None:
-        text = params.get("delta") or params.get("text") or ""
-        if not isinstance(text, str) or not text:
-            return
-        fn = getattr(agent, "_fire_reasoning_delta", None)
-        if fn is None:
-            return
-        try:
-            fn(text)
-        except Exception:
-            logger.debug("_fire_reasoning_delta raised", exc_info=True)
-
-    def _fire_agent_message_completed(item: dict) -> None:
-        text = item.get("text") or ""
-        if not isinstance(text, str) or not text.strip():
-            return
-        # display.show_commentary=false — mid-turn narration stays off the
-        # visible interim path on this runtime too (same contract as the
-        # codex_responses commentary channel).
-        if not getattr(agent, "show_commentary", True):
-            return
-        emit = getattr(agent, "_emit_interim_assistant_message", None)
-        if emit is None:
-            return
-        try:
-            emit({"role": "assistant", "content": text})
-        except Exception:
-            logger.debug(
-                "_emit_interim_assistant_message raised", exc_info=True,
-            )
-
-    def on_event(note: dict) -> None:
-        if not isinstance(note, dict):
-            return
-        method = note.get("method") or ""
-        params = note.get("params") or {}
-        if not isinstance(params, dict):
-            params = {}
-        if method == "item/agentMessage/delta":
-            _fire_text_delta(params)
-            return
-        if method in {"item/reasoning/delta", "item/reasoning/summaryDelta"}:
-            _fire_reasoning_delta(params)
-            return
-        item = params.get("item")
-        if not isinstance(item, dict):
-            return
-        item_type = item.get("type") or ""
-        if method == "item/started" and item_type in _CODEX_TOOL_ITEM_TYPES:
-            _fire_tool_started(item)
-            return
-        if method == "item/completed":
-            if item_type in _CODEX_TOOL_ITEM_TYPES:
-                _fire_tool_completed(item)
-            elif item_type == "agentMessage":
-                _fire_agent_message_completed(item)
-
-    return on_event
-
-
 def run_codex_app_server_turn(
     agent,
     *,
@@ -628,10 +244,7 @@ def run_codex_app_server_turn(
     Called from run_conversation() when agent.api_mode == "codex_app_server".
     Returns the same dict shape as the chat_completions path.
     """
-    from agent.transports.codex_app_server_session import (
-        CodexAppServerSession,
-        _ServerRequestRouting,
-    )
+    from agent.transports.codex_app_server_session import CodexAppServerSession
 
     # Lazy session: one CodexAppServerSession per AIAgent instance.
     # Spawned on first turn, reused across turns, closed at AIAgent
@@ -649,42 +262,26 @@ def run_codex_app_server_turn(
         except Exception:
             approval_callback = None
 
-        # Gateway / cron contexts have no UI to surface codex's approval
-        # requests through, so codex app-server exec / apply_patch requests
-        # fail closed (silently decline) by default. When the user has
-        # explicitly opted out of Hermes approvals — via `approvals.mode: off`
-        # in config, the /yolo session toggle, or --yolo / HERMES_YOLO_MODE —
-        # honor that and let codex's own sandbox permission profile
-        # (~/.codex/config.toml) be the policy gate instead of double-gating
-        # with a missing Hermes UI. Defaults (manual/smart/unset) preserve the
-        # current fail-closed behavior — this is a no-op for those users.
-        auto_approve_requests = False
-        try:
-            from tools.approval import is_approval_bypass_active
+        def _on_codex_event(note: dict) -> None:
+            # Bridge Codex app-server item/started notifications to Hermes
+            # tool-progress so gateways show verbose "running X" breadcrumbs
+            # on this route too (#38835).
+            progress_callback = getattr(agent, "tool_progress_callback", None)
+            if progress_callback is None:
+                return
+            mapped = _codex_note_to_tool_progress(note)
+            if mapped is None:
+                return
+            tool_name, preview, args = mapped
+            try:
+                progress_callback("tool.started", tool_name, preview, args)
+            except Exception:
+                logger.debug("codex tool-progress callback raised", exc_info=True)
 
-            auto_approve_requests = is_approval_bypass_active()
-        except Exception:
-            logger.debug(
-                "codex app-server: approval-bypass lookup failed; "
-                "keeping fail-closed default",
-                exc_info=True,
-            )
-
-        # Bridge codex JSON-RPC notifications (item/started, item/completed,
-        # item/agentMessage/delta, ...) into Hermes' gateway UI callbacks
-        # (tool_progress_callback, _fire_stream_delta,
-        # _emit_interim_assistant_message). Without this, Discord/Telegram
-        # users see no live tool-progress or interim commentary while
-        # codex_app_server is running — only the final answer (#33200).
-        # Supersedes the narrower item/started-only bridge from #38835.
         agent._codex_session = CodexAppServerSession(
             cwd=cwd,
             approval_callback=approval_callback,
-            request_routing=_ServerRequestRouting(
-                auto_approve_exec=auto_approve_requests,
-                auto_approve_apply_patch=auto_approve_requests,
-            ),
-            on_event=make_codex_app_server_event_bridge(agent),
+            on_event=_on_codex_event,
         )
 
     # NOTE: the user message is ALREADY appended to messages by the
@@ -702,16 +299,6 @@ def run_codex_app_server_turn(
         except Exception:
             pass
         agent._codex_session = None
-        _user_interrupted = bool(
-            getattr(agent, "_interrupt_requested", False)
-        )
-        _interrupt_message = (
-            getattr(agent, "_interrupt_message", None)
-            if _user_interrupted
-            else None
-        )
-        if _user_interrupted:
-            agent.clear_interrupt()
         return {
             "final_response": (
                 f"Codex app-server turn failed: {exc}. "
@@ -721,27 +308,9 @@ def run_codex_app_server_turn(
             "api_calls": 0,
             "completed": False,
             "partial": True,
-            "interrupted": _user_interrupted,
-            **(
-                {"interrupt_message": _interrupt_message}
-                if _interrupt_message
-                else {}
-            ),
             "error": str(exc),
         }
 
-    # This runtime bypasses the normal conversation-loop finalizer. Mirror its
-    # interrupt handoff/cleanup so a hard stop cannot poison the next turn and a
-    # message-bearing compatibility interrupt can still be replayed by callers.
-    _user_interrupted = bool(
-        turn.interrupted and getattr(agent, "_interrupt_requested", False)
-    )
-    _interrupt_message = (
-        getattr(agent, "_interrupt_message", None) if _user_interrupted else None
-    )
-    if _user_interrupted:
-        agent.clear_interrupt()
-
     # If the turn signalled the underlying client is wedged (deadline
     # blown, post-tool watchdog tripped, OAuth refresh died, subprocess
     # exited), retire the session so the next turn respawns codex
@@ -764,28 +333,6 @@ def run_codex_app_server_turn(
     if turn.projected_messages:
         messages.extend(turn.projected_messages)
 
-        # Persist the newly-projected assistant/tool messages ourselves.
-        # This path is an early return that bypasses conversation_loop, whose
-        # normal per-step _persist_session() calls would otherwise flush them.
-        # The inbound user turn was already flushed at turn start
-        # (turn_context.py _persist_session), and _flush_messages_to_session_db
-        # is idempotent via the intrinsic _DB_PERSISTED_MARKER — so this writes
-        # ONLY the new codex projected rows and does NOT re-write the user turn.
-        # Keeping the agent as the sole persister lets us return
-        # agent_persisted=True below, so the gateway skips its own DB write and
-        # we avoid the #860/#42039 duplicate user-message write (append_message
-        # is a raw INSERT with no dedup, so a gateway re-write would duplicate
-        # the already-flushed user turn). See gateway/run.py agent_persisted.
-        if getattr(agent, "_session_db", None) is not None:
-            try:
-                agent._flush_messages_to_session_db(messages)
-            except Exception:
-                logger.debug(
-                    "codex app-server projected-message flush failed",
-                    exc_info=True,
-                )
-
-
     # Counter ticks for the agent-improvement loop.
     # _turns_since_memory and _user_turn_count are ALREADY incremented
     # in the run_conversation() pre-loop block (lines ~11793-11817) so we
@@ -796,7 +343,6 @@ def run_codex_app_server_turn(
     agent._iters_since_skill = (
         getattr(agent, "_iters_since_skill", 0) + turn.tool_iterations
     )
-    _record_codex_app_server_compaction(agent, turn)
     usage_result = _record_codex_app_server_usage(agent, turn)
     api_calls = 1
 
@@ -847,25 +393,7 @@ def run_codex_app_server_turn(
         "api_calls": api_calls,
         "completed": not turn.interrupted and turn.error is None,
         "partial": turn.interrupted or turn.error is not None,
-        "interrupted": _user_interrupted,
-        **(
-            {"interrupt_message": _interrupt_message}
-            if _interrupt_message
-            else {}
-        ),
         "error": turn.error,
-        # The codex app-server runtime IS an early-return path that bypasses
-        # conversation_loop, but we flush the projected assistant/tool messages
-        # ourselves above (see the _flush_messages_to_session_db call after
-        # messages.extend). The inbound user turn was already flushed at turn
-        # start (turn_context._persist_session) and the flush dedups via
-        # _DB_PERSISTED_MARKER, so state.db ends up with each real message
-        # exactly once and session_search / conversation-distill see the full
-        # gateway conversation. Report agent_persisted=True so the gateway
-        # skips its own append_to_transcript DB write — writing again there
-        # would re-INSERT the already-flushed user turn (append_message has no
-        # dedup), reintroducing the #860 / #42039 duplicate-write bug.
-        "agent_persisted": True,
         "codex_thread_id": turn.thread_id,
         "codex_turn_id": turn.turn_id,
         **usage_result,
@@ -910,48 +438,18 @@ def _event_field(event: Any, name: str, default: Any = None) -> Any:
     return value if value is not None else default
 
 
-def _item_field(item: Any, name: str, default: Any = None) -> Any:
-    """Field access for nested Response items (attr-style SDK object or dict)."""
-    value = getattr(item, name, None)
-    if value is None and isinstance(item, dict):
-        value = item.get(name, default)
-    return value if value is not None else default
-
-
 def _raise_stream_error(event: Any) -> None:
     """Raise a ``_StreamErrorEvent`` from a ``type=error`` SSE frame.
 
-    The Responses spec puts the failure details at the top level of the
-    frame (``{"type": "error", "code": ..., "message": ..., "param": ...}``),
-    but the official OpenAI SDK and several OpenAI-compatible proxies wrap
-    them in an HTTP-style nested envelope instead
-    (``{"type": "error", "error": {"code": ..., "message": ..., "param": ...}}``).
-    Read the top-level fields first, then fall back to the nested envelope so
-    the error classifier sees the provider's real code/message (rate-limit vs
-    context-overflow vs entitlement) rather than the generic placeholder.
-    Port of anomalyco/opencode#36130.
-
     Imported lazily so this module stays importable from places that don't
     pull in ``run_agent`` (e.g. plugin code, doc tools).
     """
     from run_agent import _StreamErrorEvent
-
-    nested = _event_field(event, "error")
-
-    def _error_field(name: str) -> Any:
-        value = _event_field(event, name)
-        if value is None and nested is not None:
-            value = _item_field(nested, name)
-        return value
-
-    raw_message = _error_field("message")
-    if raw_message is not None and not isinstance(raw_message, str):
-        raw_message = str(raw_message)
-    message = (raw_message or "stream emitted error event").strip() or "stream emitted error event"
+    message = (_event_field(event, "message", "") or "stream emitted error event").strip()
     raise _StreamErrorEvent(
         message,
-        code=_error_field("code"),
-        param=_error_field("param"),
+        code=_event_field(event, "code"),
+        param=_event_field(event, "param"),
     )
 
 
@@ -961,7 +459,6 @@ def _consume_codex_event_stream(
     model: str,
     on_text_delta=None,
     on_reasoning_delta=None,
-    on_commentary_message=None,
     on_first_delta=None,
     on_event=None,
     interrupt_check=None,
@@ -993,11 +490,7 @@ def _consume_codex_event_stream(
     * ``on_text_delta(str)`` — fires per ``response.output_text.delta``, suppressed
       once a function_call event is seen (so tool-call turns don't bleed text
       into the chat).
-    * ``on_reasoning_delta(str)`` — fires per ``response.reasoning.*.delta`` and
-      ``phase=analysis`` message deltas. When no dedicated commentary callback
-      is supplied, commentary also uses this legacy fallback.
-    * ``on_commentary_message(str)`` — fires once per completed
-      ``phase=commentary`` message, before any following tool item executes.
+    * ``on_reasoning_delta(str)`` — fires per ``response.reasoning.*.delta``.
     * ``on_first_delta()`` — one-shot, fires on the first text delta only.
     * ``on_event(event)`` — fires for every event before any other processing.
       Used for watchdog activity, debug logging, anything wire-shape-agnostic.
@@ -1007,8 +500,6 @@ def _consume_codex_event_stream(
     collected_text_deltas: List[str] = []
     has_tool_calls = False
     first_delta_fired = False
-    active_message_phase: str | None = None
-    commentary_text_deltas: List[str] = []
     terminal_status: str = "completed"
     terminal_usage: Any = None
     terminal_response_id: str = None
@@ -1042,43 +533,9 @@ def _consume_codex_event_stream(
         if event_type == "error":
             _raise_stream_error(event)
 
-        # Track the phase of the active streamed message item.  Codex/Harmony
-        # ``commentary``/``analysis`` text is mid-turn preamble/progress
-        # narration, never the final answer.  We still collect completed output
-        # items for replay, but route those deltas to the reasoning callback so
-        # they display like thinking text instead of assistant content.
-        if event_type == "response.output_item.added":
-            item = _event_field(event, "item")
-            item_type = _item_field(item, "type", "")
-            if item_type == "message":
-                phase = _item_field(item, "phase", None)
-                active_message_phase = phase.strip().lower() if isinstance(phase, str) else None
-                if active_message_phase == "commentary":
-                    commentary_text_deltas = []
-            else:
-                active_message_phase = None
-            if "function_call" in str(item_type):
-                has_tool_calls = True
-            continue
-
         if "output_text.delta" in event_type or event_type == "response.output_text.delta":
             delta_text = _event_field(event, "delta", "")
-            if delta_text and active_message_phase == "commentary":
-                commentary_text_deltas.append(delta_text)
-                # Preserve CLI/backward compatibility when no first-class
-                # commentary consumer is installed.
-                if on_commentary_message is None and on_reasoning_delta is not None:
-                    try:
-                        on_reasoning_delta(delta_text)
-                    except Exception:
-                        logger.debug("Codex stream on_reasoning_delta raised", exc_info=True)
-            elif delta_text and active_message_phase == "analysis":
-                if on_reasoning_delta is not None:
-                    try:
-                        on_reasoning_delta(delta_text)
-                    except Exception:
-                        logger.debug("Codex stream on_reasoning_delta raised", exc_info=True)
-            elif delta_text:
+            if delta_text:
                 collected_text_deltas.append(delta_text)
                 if not has_tool_calls:
                     if not first_delta_fired:
@@ -1112,27 +569,6 @@ def _consume_codex_event_stream(
             done_item = _event_field(event, "item")
             if done_item is not None:
                 collected_output_items.append(done_item)
-                done_phase = _item_field(done_item, "phase", None)
-                done_phase = done_phase.strip().lower() if isinstance(done_phase, str) else None
-                if done_phase == "commentary" and on_commentary_message is not None:
-                    commentary_text = "".join(commentary_text_deltas).strip()
-                    if not commentary_text:
-                        content_parts = _item_field(done_item, "content", [])
-                        if isinstance(content_parts, list):
-                            commentary_text = "".join(
-                                str(_item_field(part, "text", "") or "")
-                                for part in content_parts
-                                if _item_field(part, "type", "") == "output_text"
-                            ).strip()
-                    if commentary_text:
-                        try:
-                            on_commentary_message(commentary_text)
-                        except Exception:
-                            logger.debug(
-                                "Codex stream on_commentary_message raised",
-                                exc_info=True,
-                            )
-                    commentary_text_deltas = []
             continue
 
         if event_type in _TERMINAL_EVENT_TYPES:
@@ -1233,14 +669,14 @@ def run_codex_stream(agent, api_kwargs: dict, client: Any = None, on_first_delta
     def _on_reasoning_delta(text: str) -> None:
         agent._fire_reasoning_delta(text)
 
-    def _on_commentary_message(text: str) -> None:
-        agent._fire_streamed_codex_commentary(text)
-
     def _on_event(event: Any) -> None:
         # TTFB watchdog and activity touch — runs once per SSE event.
         agent._codex_stream_last_event_ts = time.time()
         agent._touch_activity("receiving stream response")
 
+    def _interrupt_check() -> bool:
+        return bool(agent._interrupt_requested)
+
     for attempt in range(max_stream_retries + 1):
         if agent._interrupt_requested:
             raise InterruptedError("Agent interrupted before Codex stream retry")
@@ -1260,27 +696,6 @@ def run_codex_stream(agent, api_kwargs: dict, client: Any = None, on_first_delta
                 continue
             raise
 
-        # Claim the delta sink for THIS attempt (#65991) — parity with the
-        # chat_completions/anthropic/bedrock paths. If a prior attempt's
-        # stream is somehow still alive, this claim supersedes it so its
-        # late deltas are fenced out of the turn; conversely, a newer
-        # attempt supersedes us and the interrupt_check below stops our
-        # consumption immediately.
-        _writer_token = claim_stream_writer(agent)
-
-        def _interrupt_or_superseded(_tok=_writer_token) -> bool:
-            if agent._interrupt_requested:
-                return True
-            if not stream_writer_is_current(agent, _tok):
-                logger.warning(
-                    "Codex streaming attempt superseded by a newer stream; "
-                    "stopping consumption to preserve the single-writer "
-                    "invariant (model=%s).",
-                    api_kwargs.get("model", "unknown"),
-                )
-                return True
-            return False
-
         try:
             # Compatibility: some mocks/providers return a concrete response
             # instead of an iterable.  Pass it straight through.
@@ -1293,17 +708,9 @@ def run_codex_stream(agent, api_kwargs: dict, client: Any = None, on_first_delta
                     model=api_kwargs.get("model"),
                     on_text_delta=_on_text_delta,
                     on_reasoning_delta=_on_reasoning_delta,
-                    on_commentary_message=(
-                        _on_commentary_message
-                        if (
-                            getattr(agent, "interim_assistant_callback", None) is not None
-                            and getattr(agent, "show_commentary", True)
-                        )
-                        else None
-                    ),
                     on_first_delta=on_first_delta,
                     on_event=_on_event,
-                    interrupt_check=_interrupt_or_superseded,
+                    interrupt_check=_interrupt_check,
                 )
             except (_httpx.RemoteProtocolError, _httpx.ReadTimeout, _httpx.ConnectError, ConnectionError) as exc:
                 if attempt < max_stream_retries:
@@ -1352,5 +759,4 @@ __all__ = [
     "run_codex_stream",
     "run_codex_create_stream_fallback",
     "_consume_codex_event_stream",
-    "make_codex_app_server_event_bridge",
 ]

agent/coding_context.py

@@ -55,13 +55,11 @@ import json
 import logging
 import os
 import re
-import tempfile
+import subprocess
 from dataclasses import dataclass
 from pathlib import Path
 from typing import Any, Optional
 
-from hermes_cli._subprocess_compat import bounded_git_probe
-
 logger = logging.getLogger("hermes.coding_context")
 
 CODING_TOOLSET = "coding"
@@ -85,59 +83,6 @@ _PROJECT_MARKERS = (
 # Agent-instruction files surfaced separately from manifests in the snapshot.
 _CONTEXT_FILES = ("AGENTS.md", "CLAUDE.md", ".cursorrules")
 
-# Source-file extensions that make a git repo a *code* workspace even with no
-# manifest. Without this, `git init` on a notes/writing/research folder (a huge
-# non-coding use case) would flip the whole session into the coding posture just
-# for having a `.git`. A manifest still wins on its own (see `_PROJECT_MARKERS`).
-_CODE_EXTENSIONS = frozenset({
-    ".py", ".pyi", ".ipynb", ".js", ".jsx", ".ts", ".tsx", ".mjs", ".cjs",
-    ".go", ".rs", ".java", ".kt", ".kts", ".scala", ".rb", ".php", ".c", ".h",
-    ".cc", ".cpp", ".hpp", ".cs", ".swift", ".m", ".mm", ".dart", ".ex", ".exs",
-    ".lua", ".sh", ".bash", ".zsh", ".sql", ".vue", ".svelte", ".r", ".jl",
-    ".hs", ".clj", ".erl", ".pl",
-})
-
-# Dirs never worth scanning for the code check (deps/build/vcs/venv noise).
-_CODE_SCAN_SKIP_DIRS = frozenset({
-    ".git", "node_modules", "venv", ".venv", "__pycache__", "dist", "build",
-    "target", ".next", ".turbo", "vendor",
-})
-
-# Bounded sweep: a code workspace reveals itself in the first handful of entries.
-_CODE_SCAN_MAX_ENTRIES = 500
-
-
-def _has_code_files(root: Path) -> bool:
-    """Cheap, bounded check for source files in a repo's top two levels.
-
-    Lets a git repo of loose scripts (no manifest) still read as a code
-    workspace while a bare notes/writing repo does not. Scans the root and its
-    immediate subdirectories only, capped at ``_CODE_SCAN_MAX_ENTRIES`` stats —
-    a handful of readdirs at session start, not a full walk.
-    """
-    seen = 0
-    stack = [(root, True)]
-    while stack:
-        directory, is_root = stack.pop()
-        try:
-            with os.scandir(directory) as entries:
-                for entry in entries:
-                    seen += 1
-                    if seen > _CODE_SCAN_MAX_ENTRIES:
-                        return False
-                    name = entry.name
-                    try:
-                        if entry.is_file():
-                            if os.path.splitext(name)[1].lower() in _CODE_EXTENSIONS:
-                                return True
-                        elif is_root and entry.is_dir() and name not in _CODE_SCAN_SKIP_DIRS and not name.startswith("."):
-                            stack.append((Path(entry.path), False))
-                    except OSError:
-                        continue
-        except OSError:
-            continue
-    return False
-
 # Lockfile → package manager, checked in priority order.
 _PY_LOCKFILES = (("uv.lock", "uv"), ("poetry.lock", "poetry"), ("Pipfile.lock", "pipenv"))
 _JS_LOCKFILES = (
@@ -353,29 +298,6 @@ def _coding_mode(config: Optional[dict[str, Any]]) -> str:
     return "auto"
 
 
-def _coding_instructions(config: Optional[dict[str, Any]]) -> str:
-    """Standing operator instructions for the coding posture (config).
-
-    ``agent.coding_instructions`` — a string or list of strings appended to the
-    coding brief as an extra stable system block, so a user can pin project-wide
-    coding-workflow rules (e.g. "for UI work don't run tsc/lint until I approve;
-    clean the diff before committing") without editing the shipped brief.
-    Cache-safe: resolved once per session into the stable system-prompt tier,
-    like the rest of the posture.
-    """
-    if config is None:
-        try:
-            from hermes_cli.config import load_config
-
-            config = load_config()
-        except Exception:
-            config = {}
-    raw = ((config or {}).get("agent", {}) or {}).get("coding_instructions", "")
-    if isinstance(raw, (list, tuple)):
-        return "\n".join(str(item).strip() for item in raw if str(item).strip())
-    return str(raw or "").strip()
-
-
 def _resolve_cwd(cwd: Optional[str | Path]) -> Path:
     if cwd:
         return Path(cwd).expanduser()
@@ -412,18 +334,10 @@ def _marker_root(cwd: Path) -> Optional[Path]:
     """
     current = cwd.resolve()
     home = _home()
-    # Shared world-writable temp roots are never project roots: a stray
-    # manifest in /tmp (left by any process) must not flip every session
-    # whose cwd lives under the temp dir into the coding posture. Same
-    # reasoning as the $HOME skip below.
-    try:
-        temp_root = Path(tempfile.gettempdir()).resolve()
-    except Exception:
-        temp_root = None
     for depth, parent in enumerate([current, *current.parents]):
         if depth > 6:
             break
-        if parent == home or (temp_root is not None and parent == temp_root):
+        if parent == home:
             continue
         for marker in _PROJECT_MARKERS:
             if (parent / marker).exists():
@@ -454,16 +368,10 @@ def _detect_profile_name(mode: str, platform: str, cwd_str: str) -> str:
     if platform and platform.strip().lower() not in INTERACTIVE_CODING_PLATFORMS:
         return GENERAL_PROFILE.name
     cwd = Path(cwd_str)
-    # A recognized project root (manifest / AGENTS.md / .cursorrules) is a code
-    # workspace on its own — cheap stat checks, no scan.
-    if _marker_root(cwd) is not None:
-        return CODING_PROFILE.name
     git_root = _git_root(cwd)
     if git_root is not None and git_root == _home():
         git_root = None  # dotfiles repo at $HOME — not a code workspace
-    # A bare git repo only counts when it actually holds code, so `git init` on a
-    # notes/writing/research folder stays in the general posture.
-    if git_root is not None and _has_code_files(git_root):
+    if git_root is not None or _marker_root(cwd) is not None:
         return CODING_PROFILE.name
     return GENERAL_PROFILE.name
 
@@ -490,9 +398,6 @@ class RuntimeMode:
     # only to steer edit-format guidance toward the model's family — see
     # ``_edit_format_line``. Fixed for the session, so cache-safe.
     model: Optional[str] = None
-    # Standing operator instructions (``agent.coding_instructions``), appended
-    # as an extra stable system block. Empty unless the user configures it.
-    instructions: str = ""
 
     @property
     def kind(self) -> str:
@@ -520,46 +425,26 @@ class RuntimeMode:
             return None
         return [self.profile.toolset, *_enabled_mcp_servers(config)]
 
-    def system_prompt_parts(self) -> tuple[list[str], list[str], list[str]]:
-        """Return prefix, workspace, and trailing posture blocks separately.
+    def system_blocks(self) -> list[str]:
+        """Stable system-prompt blocks for this posture (brief + workspace).
 
         The operating brief carries a model-family edit-format nudge appended
         to it (one cached string, not a separate block) so the model is steered
         toward the `patch` mode it handles best — see ``_edit_format_line``.
-
-        The three lists preserve the historical flat prompt order: the brief,
-        the live workspace snapshot, then configured operator instructions.
-        Prompt assembly can therefore put a cache boundary before the snapshot
-        without changing the persisted system-prompt bytes.
         """
         if not self.is_coding:
-            return [], [], []
-        prefix: list[str] = []
-        workspace_parts: list[str] = []
-        trailing: list[str] = []
+            return []
+        blocks: list[str] = []
         if self.profile.guidance:
             brief = self.profile.guidance
             edit_line = _edit_format_line(self.model)
             if edit_line:
                 brief = f"{brief}\n{edit_line}"
-            prefix.append(brief)
+            blocks.append(brief)
         workspace = build_coding_workspace_block(self.cwd)
         if workspace:
-            workspace_parts.append(workspace)
-        # Operator instructions ride their own block so the brief (block 0) stays
-        # byte-stable and cache-keyed independently of user config.
-        if self.instructions:
-            trailing.append(f"Operator instructions (from config):\n{self.instructions}")
-        return prefix, workspace_parts, trailing
-
-    def system_blocks(self) -> list[str]:
-        """Return posture blocks in their historical display order.
-
-        ``system_prompt_parts`` is the cache-aware API. This compatibility
-        helper retains the public flat list for callers outside prompt assembly.
-        """
-        prefix, workspace, trailing = self.system_prompt_parts()
-        return [*prefix, *workspace, *trailing]
+            blocks.append(workspace)
+        return blocks
 
     def compact_skill_categories(self) -> frozenset[str]:
         """Skill categories to demote to names-only in the prompt's skill index.
@@ -611,7 +496,6 @@ def resolve_runtime_mode(
         cwd=resolved_cwd,
         config_mode=mode,
         model=model,
-        instructions=_coding_instructions(config),
     )
 
 
@@ -660,19 +544,6 @@ def coding_system_blocks(
     ).system_blocks()
 
 
-def coding_system_prompt_parts(
-    *,
-    platform: Optional[str] = None,
-    cwd: Optional[str | Path] = None,
-    config: Optional[dict[str, Any]] = None,
-    model: Optional[str] = None,
-) -> tuple[list[str], list[str], list[str]]:
-    """Return coding prefix, workspace snapshot, and trailing guidance."""
-    return resolve_runtime_mode(
-        platform=platform, cwd=cwd, config=config, model=model
-    ).system_prompt_parts()
-
-
 def coding_compact_skill_categories(
     *,
     platform: Optional[str] = None,
@@ -717,14 +588,16 @@ def _enabled_mcp_servers(config: Optional[dict[str, Any]]) -> list[str]:
 
 
 def _git(cwd: Path, *args: str) -> str:
-    """``git -C <cwd> <args>`` → stripped stdout, or ``""`` on any failure.
-
-    Uses the shared :func:`bounded_git_probe` so the post-kill cleanup is bounded
-    on Windows — a plain ``subprocess.run(timeout=...)`` here deadlocked the agent
-    turn inside ``build_coding_workspace_block`` when a killed git left a suspended
-    descendant holding the pipe handles (issue #66037).
-    """
-    return bounded_git_probe(["git", "-C", str(cwd), *args], timeout=_GIT_TIMEOUT)
+    try:
+        out = subprocess.run(
+            ["git", "-C", str(cwd), *args],
+            capture_output=True,
+            text=True,
+            timeout=_GIT_TIMEOUT,
+        )
+    except (OSError, subprocess.SubprocessError):
+        return ""
+    return out.stdout.strip() if out.returncode == 0 else ""
 
 
 def _parse_status(porcelain: str) -> tuple[dict[str, str], dict[str, int]]:
@@ -762,32 +635,25 @@ def _read_small(path: Path) -> str:
         return ""
 
 
-@dataclass(frozen=True)
-class ProjectFacts:
-    """Structured project facts — the model's verify loop, detected once.
+def _project_facts(root: Path) -> list[str]:
+    """Detected project facts for the workspace snapshot.
 
-    The same data that feeds the workspace snapshot, exposed structurally so
-    non-prompt consumers (e.g. the desktop verify UI) read it instead of
-    re-detecting and drifting from the prompt.
+    The point is to hand the model its *verify loop* up front — which manifest,
+    which package manager, and the exact test/lint/build commands — instead of
+    making it rediscover them every session. Cheap: stat calls plus reads of a
+    couple of small files; built once at prompt-build time (cache-safe).
     """
+    facts: list[str] = []
 
-    manifests: list[str]
-    package_managers: list[str]
-    verify_commands: list[str]
-    context_files: list[str]
-
-
-def detect_project_facts(root: Path) -> ProjectFacts:
-    """Detect manifests, package manager(s), verify commands, and context files.
-
-    Cheap: stat calls plus reads of a couple of small files. The single source
-    of truth for both the prompt snapshot (:func:`_project_facts`) and the
-    gateway's ``project.facts`` — so the UI never re-sniffs verify commands.
-    """
     manifests = [m for m in _PROJECT_MARKERS if m not in _CONTEXT_FILES and (root / m).is_file()]
-    package_managers = list(
-        dict.fromkeys(pm for lock, pm in (*_PY_LOCKFILES, *_JS_LOCKFILES) if (root / lock).is_file())
-    )
+    package_managers = [
+        pm for lock, pm in (*_PY_LOCKFILES, *_JS_LOCKFILES) if (root / lock).is_file()
+    ]
+    if manifests:
+        line = f"- Project: {', '.join(manifests[:6])}"
+        if package_managers:
+            line += f" ({'/'.join(dict.fromkeys(package_managers))})"
+        facts.append(line)
 
     verify: list[str] = []
     if (root / "scripts" / "run_tests.sh").is_file():
@@ -807,61 +673,17 @@ def detect_project_facts(root: Path) -> ProjectFacts:
             f"make {name}" for name in _VERIFY_TARGETS
             if re.search(rf"^{re.escape(name)}\s*:", makefile, re.MULTILINE)
         )
+    if verify:
+        deduped = list(dict.fromkeys(verify))[:_MAX_VERIFY_COMMANDS]
+        facts.append(f"- Verify: {'; '.join(deduped)}")
 
-    return ProjectFacts(
-        manifests=manifests,
-        package_managers=package_managers,
-        verify_commands=list(dict.fromkeys(verify))[:_MAX_VERIFY_COMMANDS],
-        context_files=[c for c in _CONTEXT_FILES if (root / c).is_file()],
-    )
-
-
-def _project_facts(root: Path) -> list[str]:
-    """Render :func:`detect_project_facts` as workspace-snapshot lines.
-
-    Hands the model its *verify loop* up front — which manifest, which package
-    manager, and the exact test/lint/build commands — instead of making it
-    rediscover them every session. Built once at prompt-build time; the string
-    output must stay byte-stable to preserve the prompt cache.
-    """
-    f = detect_project_facts(root)
-    facts: list[str] = []
-
-    if f.manifests:
-        line = f"- Project: {', '.join(f.manifests[:6])}"
-        if f.package_managers:
-            line += f" ({'/'.join(f.package_managers)})"
-        facts.append(line)
-    if f.verify_commands:
-        facts.append(f"- Verify: {'; '.join(f.verify_commands)}")
-    if f.context_files:
-        facts.append(f"- Context files: {', '.join(f.context_files)}")
+    context_files = [c for c in _CONTEXT_FILES if (root / c).is_file()]
+    if context_files:
+        facts.append(f"- Context files: {', '.join(context_files)}")
 
     return facts
 
 
-def project_facts_for(cwd: Optional[str | Path] = None) -> Optional[dict[str, Any]]:
-    """Structured project facts for ``cwd`` — ``None`` outside a workspace.
-
-    Same detection the system-prompt snapshot uses (git root, else marker root),
-    exposed for non-prompt consumers (the desktop verify UI) so they never
-    re-derive "are we coding?" or duplicate the verify-command sniffing.
-    """
-    resolved = _resolve_cwd(cwd)
-    root = _git_root(resolved) or _marker_root(resolved)
-    if root is None:
-        return None
-
-    f = detect_project_facts(root)
-    return {
-        "root": str(root),
-        "manifests": f.manifests,
-        "packageManagers": f.package_managers,
-        "verifyCommands": f.verify_commands,
-        "contextFiles": f.context_files,
-    }
-
-
 def build_coding_workspace_block(cwd: Optional[str | Path] = None) -> str:
     """Workspace snapshot for the system prompt (empty outside a workspace).
 

agent/context_breakdown.py

@@ -1,156 +0,0 @@
-"""Live session context-window breakdown for UI surfaces.
-
-Estimates how the next provider request is composed: system prompt tiers,
-tool schemas, and conversation history. Uses the same rough char/4 heuristic
-as ``agent.model_metadata.estimate_request_tokens_rough`` so numbers align
-with compression thresholds — not exact tokenizer counts.
-"""
-
-from __future__ import annotations
-
-import json
-import re
-from typing import Any, Dict, List, Optional, Sequence, Tuple
-
-_SKILLS_BLOCK_RE = re.compile(r"<available_skills>.*?</available_skills>", re.DOTALL)
-
-_SUBAGENT_TOOL_NAMES = frozenset({"delegate_task"})
-
-_CATEGORY_COLORS = {
-    "system_prompt": "var(--context-usage-system)",
-    "tool_definitions": "var(--context-usage-tools)",
-    "rules": "var(--context-usage-rules)",
-    "skills": "var(--context-usage-skills)",
-    "mcp": "var(--context-usage-mcp)",
-    "subagent_definitions": "var(--context-usage-subagents)",
-    "memory": "var(--context-usage-memory)",
-    "conversation": "var(--context-usage-conversation)",
-}
-
-
-def _chars_to_tokens(text: str) -> int:
-    if not text:
-        return 0
-    return (len(text) + 3) // 4
-
-
-def _json_tokens(value: Any) -> int:
-    if not value:
-        return 0
-    return _chars_to_tokens(json.dumps(value, ensure_ascii=False))
-
-
-def _tool_name(tool: dict) -> str:
-    fn = tool.get("function") if isinstance(tool, dict) else None
-    if isinstance(fn, dict):
-        return str(fn.get("name") or "")
-    return str(tool.get("name") or "")
-
-
-def _split_tools(tools: Sequence[dict]) -> Tuple[List[dict], List[dict], List[dict]]:
-    builtin: List[dict] = []
-    mcp: List[dict] = []
-    subagent: List[dict] = []
-    for tool in tools:
-        name = _tool_name(tool)
-        if name.startswith("mcp_"):
-            mcp.append(tool)
-        elif name in _SUBAGENT_TOOL_NAMES:
-            subagent.append(tool)
-        else:
-            builtin.append(tool)
-    return builtin, mcp, subagent
-
-
-def _memory_blocks(agent: Any) -> Tuple[str, str]:
-    memory_block = ""
-    user_block = ""
-    store = getattr(agent, "_memory_store", None)
-    if store is None:
-        return memory_block, user_block
-    try:
-        if getattr(agent, "_memory_enabled", True):
-            memory_block = store.format_for_system_prompt("memory") or ""
-        if getattr(agent, "_user_profile_enabled", True):
-            user_block = store.format_for_system_prompt("user") or ""
-    except Exception:
-        pass
-    return memory_block, user_block
-
-
-def _strip_blocks(text: str, *blocks: str) -> str:
-    out = text
-    for block in blocks:
-        if block:
-            out = out.replace(block, "")
-    return out.strip()
-
-
-def compute_session_context_breakdown(
-    agent: Any,
-    messages: Optional[List[dict]] = None,
-) -> Dict[str, Any]:
-    """Return a Cursor-style context usage breakdown for one live agent."""
-    from agent.model_metadata import estimate_messages_tokens_rough
-    from agent.system_prompt import build_system_prompt_parts
-
-    parts = build_system_prompt_parts(agent)
-    stable = parts.get("stable", "") or ""
-    context = parts.get("context", "") or ""
-    volatile = parts.get("volatile", "") or ""
-
-    skills_match = _SKILLS_BLOCK_RE.search(stable)
-    skills_index = skills_match.group(0) if skills_match else ""
-
-    memory_block, user_block = _memory_blocks(agent)
-    memory_text = "\n\n".join(part for part in (memory_block, user_block) if part).strip()
-
-    system_core = _strip_blocks(stable, skills_index)
-    system_tail = _strip_blocks(volatile, memory_block, user_block)
-    system_prompt_text = "\n\n".join(part for part in (system_core, system_tail) if part).strip()
-
-    tools = list(getattr(agent, "tools", None) or [])
-    builtin_tools, mcp_tools, subagent_tools = _split_tools(tools)
-
-    conversation_tokens = estimate_messages_tokens_rough(messages or [])
-
-    categories = [
-        ("system_prompt", "System prompt", _chars_to_tokens(system_prompt_text)),
-        ("tool_definitions", "Tool definitions", _json_tokens(builtin_tools)),
-        ("rules", "Rules", _chars_to_tokens(context)),
-        ("skills", "Skills", _chars_to_tokens(skills_index)),
-        ("mcp", "MCP", _json_tokens(mcp_tools)),
-        ("subagent_definitions", "Subagent definitions", _json_tokens(subagent_tools)),
-        ("memory", "Memory", _chars_to_tokens(memory_text)),
-        ("conversation", "Conversation", conversation_tokens),
-    ]
-
-    estimated_total = sum(tokens for _, _, tokens in categories)
-
-    comp = getattr(agent, "context_compressor", None)
-    context_max = int(getattr(comp, "context_length", 0) or 0) if comp else 0
-    measured_used = int(getattr(comp, "last_prompt_tokens", 0) or 0) if comp else 0
-    context_used = measured_used if measured_used > 0 else estimated_total
-    context_percent = (
-        max(0, min(100, round(context_used / context_max * 100)))
-        if context_max
-        else 0
-    )
-
-    return {
-        "categories": [
-            {
-                "color": _CATEGORY_COLORS.get(category_id, "var(--ui-text-tertiary)"),
-                "id": category_id,
-                "label": label,
-                "tokens": tokens,
-            }
-            for category_id, label, tokens in categories
-            if tokens > 0
-        ],
-        "context_max": context_max,
-        "context_percent": context_percent,
-        "context_used": context_used,
-        "estimated_total": estimated_total,
-        "model": getattr(agent, "model", "") or "",
-    }

agent/context_engine.py

@@ -26,64 +26,7 @@ Lifecycle:
 """
 
 from abc import ABC, abstractmethod
-from typing import Any, Dict, List, Optional
-
-from agent.redact import redact_sensitive_text
-
-
-MEMORY_CONTEXT_MAX_CHARS = 6_000
-_MEMORY_CONTEXT_HEAD_CHARS = 4_000
-_MEMORY_CONTEXT_TAIL_CHARS = 1_500
-_MEMORY_CONTEXT_TRUNCATION_MARKER = "\n...[memory provider context truncated]...\n"
-
-
-def sanitize_memory_context(memory_context: str) -> str:
-    """Prepare provider context for a context-engine/LLM egress boundary."""
-    sanitized = redact_sensitive_text(
-        memory_context.strip(),
-        force=True,
-        redact_url_credentials=True,
-    )
-    if len(sanitized) <= MEMORY_CONTEXT_MAX_CHARS:
-        return sanitized
-    return (
-        sanitized[:_MEMORY_CONTEXT_HEAD_CHARS]
-        + _MEMORY_CONTEXT_TRUNCATION_MARKER
-        + sanitized[-_MEMORY_CONTEXT_TAIL_CHARS:]
-    )
-
-
-def automatic_compaction_status_message(
-    engine: Any,
-    *,
-    phase: str,
-    default_message: str,
-    **context: Any,
-) -> str | None:
-    """Resolve host-visible status for an automatic compaction event.
-
-    Engines can suppress routine automatic status with
-    ``emit_automatic_compaction_status = False`` or customize it by defining
-    ``get_automatic_compaction_status_message(...)``. Empty strings and
-    ``None`` mean "do not emit a lifecycle status".
-    """
-    if not getattr(engine, "emit_automatic_compaction_status", True):
-        return None
-
-    formatter = getattr(engine, "get_automatic_compaction_status_message", None)
-    if callable(formatter):
-        message = formatter(
-            phase=phase,
-            default_message=default_message,
-            **context,
-        )
-    else:
-        message = default_message
-
-    if message is None:
-        return None
-    message = str(message).strip()
-    return message or None
+from typing import Any, Dict, List
 
 
 class ContextEngine(ABC):
@@ -122,12 +65,6 @@ class ContextEngine(ABC):
     protect_first_n: int = 3
     protect_last_n: int = 6
 
-    # User-visible lifecycle status for automatic host-triggered compaction.
-    # Alternative engines that treat compaction as routine background
-    # maintenance can set this false to keep successful automatic passes silent;
-    # warnings, errors, and explicit manual commands should still surface.
-    emit_automatic_compaction_status: bool = True
-
     # -- Core interface ----------------------------------------------------
 
     @abstractmethod
@@ -146,27 +83,12 @@ class ContextEngine(ABC):
     def should_compress(self, prompt_tokens: int = None) -> bool:
         """Return True if compaction should fire this turn."""
 
-    def should_compress_info(self, prompt_tokens: int = None) -> "tuple[bool, str | None]":
-        """Return ``(should_compress, reason)``.
-
-        The base implementation is backward-compatible: engines that only
-        implement ``should_compress`` get ``(should_compress(prompt_tokens),
-        None)``. Concrete engines with richer block reasons (e.g. a
-        summary-LLM cooldown or an anti-thrashing guard) override this to
-        surface a human-readable reason so callers can warn the user instead
-        of silently skipping compression. Added for the silent-overflow
-        warning fix (#62625) so plugin engines don't raise AttributeError.
-        """
-        return self.should_compress(prompt_tokens), None
-
     @abstractmethod
     def compress(
         self,
         messages: List[Dict[str, Any]],
-        current_tokens: Optional[int] = None,
-        focus_topic: Optional[str] = None,
-        force: bool = False,
-        memory_context: str = "",
+        current_tokens: int = None,
+        focus_topic: str = None,
     ) -> List[Dict[str, Any]]:
         """Compact the message list and return the new message list.
 
@@ -181,152 +103,8 @@ class ContextEngine(ABC):
                 Engines that support guided compression should prioritise
                 preserving information related to this topic.  Engines that
                 don't support it may simply ignore this argument.
-            force: Whether a user-requested compression should bypass an
-                engine-owned cooldown. Engines without cooldowns may ignore it.
-            memory_context: Text returned by memory providers immediately before
-                compaction. Summarizing engines should include non-empty text in
-                their handoff prompt. Older engines may omit this parameter; the
-                host filters unsupported optional arguments by signature.
         """
 
-    # -- Optional: proactive tool-result prune -----------------------------
-
-    def prune_tool_results_only(
-        self,
-        messages: List[Dict[str, Any]],
-        current_tokens: int | None = None,
-    ) -> tuple[List[Dict[str, Any]], int]:
-        """Deterministically trim old tool-result payloads without an LLM call.
-
-        Runs on a low, cost-oriented trigger independent of ``should_compress``
-        so large-window engines can reclaim re-sent tool output long before full
-        compaction would fire. Returns ``(messages, n_pruned)``.
-
-        Default is a safe no-op: the list is returned unchanged with ``0``
-        pruned. Engines that don't implement a cheap prune — and any engine that
-        predates this hook — inherit this default, so the agent loop's
-        post-tool-call prune path never raises ``AttributeError`` on them. The
-        built-in ContextCompressor overrides this with the real implementation.
-        """
-        return messages, 0
-
-    # -- Optional: per-turn context selection (distinct from compression) --
-
-    def select_context(
-        self,
-        request_messages: List[Dict[str, Any]],
-        *,
-        conversation_messages: List[Dict[str, Any]] = None,
-        incoming_message: Dict[str, Any] = None,
-        budget_tokens: int = 0,
-    ) -> List[Dict[str, Any]]:
-        """Optionally choose/replace the context for THIS request, pre-generation.
-
-        Called every turn after the request message list is assembled and
-        before it is dispatched to the provider — independent of
-        ``should_compress()``. This lets an engine *select* which context
-        enters the prompt (retrieval, topic routing, role/branch switching)
-        rather than *shrink* context that is already there. The two verbs are
-        orthogonal:
-
-          - ``compress()``      : context is too long  -> make it shorter.
-          - ``select_context()``: this turn belongs to a different context
-                                  -> use that one instead.
-
-        Without this hook, engines that need per-turn access to the message
-        list have to force ``should_compress()`` to return ``True`` so that
-        ``compress()`` is invoked every turn purely as a callback — which
-        conflates selection with compression and degrades behaviour when the
-        engine's backend is unavailable. ``select_context()`` removes the need
-        for that workaround.
-
-        The returned list is request-only: it replaces the messages sent to
-        the provider for this single call and MUST NOT be treated as persisted
-        transcript state. The conversation history in the session DB is left
-        untouched, so nothing leaks across turns. Return ``None`` to leave the
-        request unchanged.
-
-        Unlike the ``pre_llm_call`` plugin hook (which appends to the user
-        message and intentionally never rewrites the list, to preserve the
-        cache prefix), ``select_context()`` may *replace* the message list.
-
-        Ordering / cache contract: the host runs this hook **before** prompt
-        cache-control and **before** every request sanitizer (orphaned-tool
-        cleanup, thinking-only/role normalization, whitespace/JSON
-        normalization). So (a) whatever the hook returns still passes through
-        the same validation as any request — a malformed replacement cannot
-        reach the provider — and (b) prompt-cache stability (an AGENTS.md
-        invariant) is preserved: the default no-op leaves the request
-        byte-identical, so cache behaviour is unchanged for the built-in
-        compressor and any non-implementing engine. An engine that *does*
-        replace the list changes its own cache prefix by definition; that is
-        the engine's concern, and cache-control breakpoints are re-derived on
-        the selected list. The hook is evaluated per provider request (so it
-        re-runs on retries within a turn), consistent with "select the context
-        for THIS request".
-
-        Args:
-            request_messages: The assembled request message list (system
-                prompt + history + any ephemeral prefill), in OpenAI format.
-            conversation_messages: The unmodified persisted conversation
-                history, for reference only (do not mutate).
-            incoming_message: The current turn's user message, if available.
-            budget_tokens: The active model's context length, or 0 if unknown.
-
-        Default returns ``None`` (no-op) — zero impact on the built-in
-        compressor or any existing engine.
-        """
-        return None
-
-    def on_turn_complete(
-        self,
-        messages: List[Dict[str, Any]],
-        usage: Dict[str, Any] = None,
-        **kwargs: Any,
-    ) -> None:
-        """Observe a finished user turn (post-turn ingestion / observation).
-
-        Called from the standard turn-finalization path once the assistant/tool
-        loop completes, with the finalized in-memory transcript snapshot. This
-        is the complement to ``select_context()``: selection happens *before*
-        the request, while observation happens *after* the turn. It lets an
-        engine ingest, index, summarize, or update routing / topic / session
-        state from what actually happened — so the next ``select_context()``
-        can act on it.
-
-        Coverage: this fires from the normal finalization seam. Some abnormal
-        early-return paths in the loop (e.g. a content-policy block or a
-        provider terminal failure) persist and return without routing through
-        finalization, and therefore do not currently emit this hook. Treat it
-        as a best-effort post-turn observation for completed turns, not a
-        guaranteed callback for every possible early exit; unifying all
-        terminal paths behind one finalization seam is a separate follow-up.
-
-        Together the two hooks remove the need to abuse ``should_compress()`` /
-        ``compress()`` as a generic per-turn callback just to observe history,
-        and they cover the case where a turn finishes and there may be no next
-        request from which to infer the previous turn.
-
-        ``messages`` is a shallow copy and should be treated as read-only:
-        return values are ignored and this hook must not rely on transcript
-        mutation for persistence. ``kwargs`` may include ``turn_id``,
-        ``task_id``, ``api_call_count``, ``interrupted``, ``failed``, and
-        ``turn_exit_reason``.
-
-        ``usage`` carries the completed turn's canonical token usage (the same
-        dict shape passed to ``update_from_response`` — ``prompt_tokens`` /
-        ``completion_tokens`` / ``total_tokens`` plus the canonical
-        ``input_tokens`` / ``output_tokens`` / ``cache_read_tokens`` /
-        ``cache_write_tokens`` / ``reasoning_tokens`` buckets) so an engine can
-        weigh how large/expensive the selected context actually was when
-        deciding the next ``select_context()``. It is ``None`` on finalized
-        turns that never reached a provider response (e.g. interrupt); engines
-        must treat it as optional.
-
-        Default is a no-op.
-        """
-        return None
-
     # -- Optional: pre-flight check ----------------------------------------
 
     def should_compress_preflight(self, messages: List[Dict[str, Any]]) -> bool:
@@ -346,27 +124,6 @@ class ContextEngine(ABC):
         """
         return False
 
-    def get_automatic_compaction_status_message(
-        self,
-        *,
-        phase: str,
-        default_message: str,
-        **context: Any,
-    ) -> str | None:
-        """Return user-visible status for automatic host-triggered compaction.
-
-        Return ``None`` to suppress successful automatic lifecycle status for
-        this compaction event. ``phase`` identifies the host call site (for
-        example ``"preflight"`` or ``"compress"``). ``context`` contains
-        best-effort fields such as ``approx_tokens`` and ``threshold_tokens``.
-
-        This hook does not control warning/error messages or explicit manual
-        commands such as ``/compress``.
-        """
-        if not self.emit_automatic_compaction_status:
-            return None
-        return default_message
-
     # -- Optional: manual /compress preflight ------------------------------
 
     def has_content_to_compress(self, messages: List[Dict[str, Any]]) -> bool:
@@ -437,17 +194,12 @@ class ContextEngine(ABC):
 
         Default returns the standard fields run_agent.py expects.
         """
-        # Clamp the -1 "compression just ran, awaiting real usage" sentinel
-        # (set by conversation_compression) to 0 so status readers don't see a
-        # raw -1 or a negative usage_percent on the transitional turn. Mirrors
-        # the CLI/gateway status-bar paths (cli.py, tui_gateway/server.py).
-        last_prompt = self.last_prompt_tokens if self.last_prompt_tokens > 0 else 0
         return {
-            "last_prompt_tokens": last_prompt,
+            "last_prompt_tokens": self.last_prompt_tokens,
             "threshold_tokens": self.threshold_tokens,
             "context_length": self.context_length,
             "usage_percent": (
-                min(100, last_prompt / self.context_length * 100)
+                min(100, self.last_prompt_tokens / self.context_length * 100)
                 if self.context_length else 0
             ),
             "compression_count": self.compression_count,
@@ -471,19 +223,4 @@ class ContextEngine(ABC):
         (e.g. recalculate DAG budgets, switch summary models).
         """
         self.context_length = context_length
-        # Apply per-model threshold overrides if set (longest substring match).
-        # Falls back to _config_threshold_percent (the raw config value) when
-        # no override matches. Plugin engines that override update_model() can
-        # call resolve_model_threshold() for the same logic.
-        from agent.context_compressor import resolve_model_threshold
-        if not hasattr(self, "_config_threshold_percent"):
-            # Snapshot the pre-override percent ONCE so repeated model
-            # switches fall back to the engine's configured value, not the
-            # previous model's override.
-            self._config_threshold_percent = self.threshold_percent
-        self._base_threshold_percent = resolve_model_threshold(
-            model, getattr(self, "model_thresholds", {}),
-            self._config_threshold_percent,
-        )
-        self.threshold_percent = self._base_threshold_percent
         self.threshold_tokens = int(context_length * self.threshold_percent)

agent/context_references.py

@@ -12,7 +12,6 @@ from pathlib import Path
 from typing import Awaitable, Callable
 
 from agent.model_metadata import estimate_tokens_rough
-from hermes_cli._subprocess_compat import IS_WINDOWS, windows_hide_flags
 
 _QUOTED_REFERENCE_VALUE = r'(?:`[^`\n]+`|"[^"\n]+"|\'[^\'\n]+\')'
 REFERENCE_PATTERN = re.compile(
@@ -152,24 +151,13 @@ async def preprocess_context_references_async(
     blocks: list[str] = []
     injected_tokens = 0
 
-    # Expand all references concurrently. Each _expand_reference is independent
-    # (no shared state during expansion) — a message with several @url: refs
-    # would otherwise pay one full web_extract round-trip per ref in series.
-    # gather preserves positional order, so we reassemble warnings/blocks in the
-    # original ref order exactly as the prior serial loop did; the token-budget
-    # check below is unchanged (it runs once, after all refs are expanded).
-    expanded = await asyncio.gather(
-        *(
-            _expand_reference(
-                ref,
-                cwd_path,
-                url_fetcher=url_fetcher,
-                allowed_root=allowed_root_path,
-            )
-            for ref in refs
+    for ref in refs:
+        warning, block = await _expand_reference(
+            ref,
+            cwd_path,
+            url_fetcher=url_fetcher,
+            allowed_root=allowed_root_path,
         )
-    )
-    for warning, block in expanded:
         if warning:
             warnings.append(warning)
         if block:
@@ -302,16 +290,14 @@ def _expand_git_reference(
     args: list[str],
     label: str,
 ) -> tuple[str | None, str | None]:
-    _popen_kwargs = {"creationflags": windows_hide_flags()} if IS_WINDOWS else {}
     try:
         result = subprocess.run(
             ["git", *args],
             cwd=cwd,
             capture_output=True,
-            text=True, encoding='utf-8', errors='replace',
+            text=True,
             timeout=30,
             stdin=subprocess.DEVNULL,
-            **_popen_kwargs,
         )
     except subprocess.TimeoutExpired:
         return f"{ref.raw}: git command timed out (30s)", None
@@ -339,9 +325,9 @@ async def _fetch_url_content(
 async def _default_url_fetcher(url: str) -> str:
     from tools.web_tools import web_extract_tool
 
-    raw = await web_extract_tool([url], format="markdown")
+    raw = await web_extract_tool([url], format="markdown", use_llm_processing=True)
     payload = json.loads(raw)
-    docs = payload.get("results", [])
+    docs = payload.get("data", {}).get("documents", [])
     if not docs:
         return ""
     doc = docs[0]
@@ -381,37 +367,6 @@ def _ensure_reference_path_allowed(path: Path) -> None:
             continue
         raise ValueError("path is a sensitive credential or internal Hermes path and cannot be attached")
 
-    # Anchor to the canonical read deny-list (agent/file_safety.get_read_block_error),
-    # the single source of truth used by the file/terminal read path. The narrow
-    # list above predates that guard and never caught the real credential stores:
-    # provider keys (auth.json), Anthropic OAuth tokens (.anthropic_oauth.json),
-    # MCP OAuth material (mcp-tokens/), webhook HMAC secrets, and project-local
-    # .env files. That gap matters because the gateway feeds UNTRUSTED remote
-    # message text into reference expansion, so `@file:~/.hermes/auth.json` from a
-    # chat peer would otherwise read the operator's keys straight into context.
-    # Routing through the canonical guard closes the gap today and keeps this path
-    # protected automatically whenever that deny-list grows.
-    try:
-        from agent.file_safety import get_read_block_error
-
-        if get_read_block_error(str(path)) is not None:
-            raise ValueError(
-                "path is a sensitive credential or internal Hermes path and cannot be attached"
-            )
-    except ValueError:
-        raise
-    except Exception:
-        # Fail CLOSED on the security path. This guard exists specifically to
-        # cover credential stores the narrow list above misses (auth.json,
-        # .anthropic_oauth.json, mcp-tokens/, ...). If the canonical lookup
-        # ever fails, silently falling through would re-open that exact hole —
-        # the gateway feeds untrusted remote text here, so a probe could then
-        # attach the operator's keys. Refuse instead: a spurious block on a
-        # legitimate file is a recoverable annoyance; a leaked credential is not.
-        raise ValueError(
-            "path could not be verified against the credential deny-list and cannot be attached"
-        )
-
 
 def _strip_trailing_punctuation(value: str) -> str:
     stripped = value.rstrip(TRAILING_PUNCTUATION)
@@ -528,16 +483,14 @@ def _iter_visible_entries(path: Path, cwd: Path, limit: int) -> list[Path]:
 
 
 def _rg_files(path: Path, cwd: Path, limit: int) -> list[Path] | None:
-    _popen_kwargs = {"creationflags": windows_hide_flags()} if IS_WINDOWS else {}
     try:
         result = subprocess.run(
             ["rg", "--files", str(path.relative_to(cwd))],
             cwd=cwd,
             capture_output=True,
-            text=True, encoding='utf-8', errors='replace',
+            text=True,
             timeout=10,
             stdin=subprocess.DEVNULL,
-            **_popen_kwargs,
         )
     except (FileNotFoundError, OSError, subprocess.TimeoutExpired):
         return None

agent/copilot_acp_client.py

@@ -21,14 +21,8 @@ from pathlib import Path
 from types import SimpleNamespace
 from typing import Any
 
-from openai.types.chat.chat_completion_message_tool_call import (
-    ChatCompletionMessageToolCall,
-    Function,
-)
-
-from agent.file_safety import get_read_block_error, get_write_denied_error
+from agent.file_safety import get_read_block_error, is_write_denied
 from agent.redact import redact_sensitive_text
-from tools.environments.local import hermes_subprocess_env
 
 ACP_MARKER_BASE_URL = "acp://copilot"
 _DEFAULT_TIMEOUT_SECONDS = 900.0
@@ -100,10 +94,7 @@ def _resolve_home_dir() -> str:
 
 
 def _build_subprocess_env() -> dict[str, str]:
-    # Copilot ACP is a model-driving CLI executor: it legitimately needs LLM
-    # provider credentials. Route through the central helper so Tier-1 secrets
-    # (gateway bot tokens, GitHub auth, infra) are still stripped (#29157).
-    env = hermes_subprocess_env(inherit_credentials=True)
+    env = os.environ.copy()
     home = _resolve_home_dir()
     env["HOME"] = home
     from hermes_constants import apply_subprocess_home_env
@@ -233,73 +224,11 @@ def _render_message_content(content: Any) -> str:
     return str(content).strip()
 
 
-def _build_openai_tool_call(
-    *,
-    call_id: str,
-    name: str,
-    arguments: str,
-) -> ChatCompletionMessageToolCall:
-    """Build an OpenAI-compatible tool-call object for downstream handling."""
-    return ChatCompletionMessageToolCall(
-        id=call_id,
-        call_id=call_id,
-        response_item_id=None,
-        type="function",
-        function=Function(name=name, arguments=arguments),
-    )
-
-
-def _completion_to_stream_chunks(completion: SimpleNamespace) -> list[SimpleNamespace]:
-    """Convert a one-shot ACP response into OpenAI-style stream chunks."""
-    choice = completion.choices[0]
-    message = choice.message
-    tool_call_deltas = None
-    if message.tool_calls:
-        tool_call_deltas = []
-        for index, tool_call in enumerate(message.tool_calls):
-            tool_call_deltas.append(
-                SimpleNamespace(
-                    index=index,
-                    id=getattr(tool_call, "id", None),
-                    type=getattr(tool_call, "type", "function"),
-                    function=SimpleNamespace(
-                        name=getattr(tool_call.function, "name", None),
-                        arguments=getattr(tool_call.function, "arguments", None),
-                    ),
-                )
-            )
-
-    delta = SimpleNamespace(
-        role="assistant",
-        content=message.content or None,
-        tool_calls=tool_call_deltas,
-        reasoning_content=message.reasoning_content,
-        reasoning=message.reasoning,
-    )
-    data_chunk = SimpleNamespace(
-        choices=[
-            SimpleNamespace(
-                index=0,
-                delta=delta,
-                finish_reason=choice.finish_reason,
-            )
-        ],
-        model=completion.model,
-        usage=None,
-    )
-    usage_chunk = SimpleNamespace(
-        choices=[],
-        model=completion.model,
-        usage=completion.usage,
-    )
-    return [data_chunk, usage_chunk]
-
-
-def _extract_tool_calls_from_text(text: str) -> tuple[list[ChatCompletionMessageToolCall], str]:
+def _extract_tool_calls_from_text(text: str) -> tuple[list[SimpleNamespace], str]:
     if not isinstance(text, str) or not text.strip():
         return [], ""
 
-    extracted: list[ChatCompletionMessageToolCall] = []
+    extracted: list[SimpleNamespace] = []
     consumed_spans: list[tuple[int, int]] = []
 
     def _try_add_tool_call(raw_json: str) -> None:
@@ -323,10 +252,12 @@ def _extract_tool_calls_from_text(text: str) -> tuple[list[ChatCompletionMessage
             call_id = f"acp_call_{len(extracted)+1}"
 
         extracted.append(
-            _build_openai_tool_call(
+            SimpleNamespace(
+                id=call_id,
                 call_id=call_id,
-                name=fn_name.strip(),
-                arguments=fn_args,
+                response_item_id=None,
+                type="function",
+                function=SimpleNamespace(name=fn_name.strip(), arguments=fn_args),
             )
         )
 
@@ -445,7 +376,6 @@ class CopilotACPClient:
         timeout: float | None = None,
         tools: list[dict[str, Any]] | None = None,
         tool_choice: Any = None,
-        stream: bool = False,
         **_: Any,
     ) -> Any:
         prompt_text = _format_messages_as_prompt(
@@ -492,31 +422,23 @@ class CopilotACPClient:
         )
         finish_reason = "tool_calls" if tool_calls else "stop"
         choice = SimpleNamespace(message=assistant_message, finish_reason=finish_reason)
-        completion = SimpleNamespace(
+        return SimpleNamespace(
             choices=[choice],
             usage=usage,
             model=model or "copilot-acp",
         )
-        if stream:
-            return _completion_to_stream_chunks(completion)
-        return completion
 
     def _run_prompt(self, prompt_text: str, *, timeout_seconds: float) -> tuple[str, str]:
         try:
-            # Hide the console the CLI child would otherwise flash on Windows
-            # (#56747). Hide-only — stdio pipes stay intact for the ACP wire.
-            from hermes_cli._subprocess_compat import windows_hide_flags
-
             proc = subprocess.Popen(
                 [self._acp_command] + self._acp_args,
                 stdin=subprocess.PIPE,
                 stdout=subprocess.PIPE,
                 stderr=subprocess.PIPE,
-                text=True, encoding='utf-8', errors='replace',
+                text=True,
                 bufsize=1,
                 cwd=self._acp_cwd,
                 env=_build_subprocess_env(),
-                creationflags=windows_hide_flags(),
             )
         except FileNotFoundError as exc:
             raise RuntimeError(
@@ -708,7 +630,7 @@ class CopilotACPClient:
                 if block_error:
                     raise PermissionError(block_error)
                 try:
-                    content = path.read_text(encoding="utf-8")
+                    content = path.read_text()
                 except FileNotFoundError:
                     content = ""
                 line = params.get("line")
@@ -732,11 +654,12 @@ class CopilotACPClient:
         elif method == "fs/write_text_file":
             try:
                 path = _ensure_path_within_cwd(str(params.get("path") or ""), cwd)
-                denied = get_write_denied_error(str(path))
-                if denied:
-                    raise PermissionError(denied)
+                if is_write_denied(str(path)):
+                    raise PermissionError(
+                        f"Write denied: '{path}' is a protected system/credential file."
+                    )
                 path.parent.mkdir(parents=True, exist_ok=True)
-                path.write_text(str(params.get("content") or ""), encoding="utf-8")
+                path.write_text(str(params.get("content") or ""))
                 response = {
                     "jsonrpc": "2.0",
                     "id": message_id,

agent/credential_persistence.py

@@ -22,7 +22,7 @@ _PERSISTABLE_PROVIDER_SOURCES = frozenset({
     ("minimax-oauth", "oauth"),
     ("nous", "device_code"),
     ("openai-codex", "device_code"),
-    ("xai-oauth", "device_code"),
+    ("xai-oauth", "loopback_pkce"),
 })
 
 _SAFE_SECRETISH_METADATA_KEYS = frozenset({

agent/credential_sources.py

@@ -164,7 +164,7 @@ def _remove_env_source(provider: str, removed) -> RemovalResult:
         if env_path.exists():
             env_in_dotenv = any(
                 line.strip().startswith(f"{env_var}=")
-                for line in env_path.read_text(errors="replace", encoding="utf-8").splitlines()
+                for line in env_path.read_text(errors="replace").splitlines()
             )
     except OSError:
         pass
@@ -265,7 +265,7 @@ def _remove_minimax_oauth(provider: str, removed) -> RemovalResult:
     return result
 
 
-def _remove_xai_oauth_device_code(provider: str, removed) -> RemovalResult:
+def _remove_xai_oauth_loopback_pkce(provider: str, removed) -> RemovalResult:
     """xAI OAuth tokens live in auth.json providers.xai-oauth — clear them.
 
     Without this step, ``hermes auth remove xai-oauth <N>`` silently undoes
@@ -275,6 +275,11 @@ def _remove_xai_oauth_device_code(provider: str, removed) -> RemovalResult:
     entry from the still-present singleton — credentials reappear with no
     user feedback. Clearing the singleton in step with the suppression set
     by the central dispatcher makes the removal stick.
+
+    Belt-and-braces against the manual entry path: ``hermes auth add
+    xai-oauth`` produces a ``manual:xai_pkce`` entry whose removal step
+    falls through to "unregistered → nothing to clean up" (correct —
+    manual entries are pool-only).
     """
     result = RemovalResult()
     if _clear_auth_store_provider(provider):
@@ -418,8 +423,8 @@ def _register_all_sources() -> None:
         description="auth.json providers.openai-codex + ~/.codex/auth.json",
     ))
     register(RemovalStep(
-        provider="xai-oauth", source_id="device_code",
-        remove_fn=_remove_xai_oauth_device_code,
+        provider="xai-oauth", source_id="loopback_pkce",
+        remove_fn=_remove_xai_oauth_loopback_pkce,
         description="auth.json providers.xai-oauth",
     ))
     register(RemovalStep(

agent/credits_tracker.py

@@ -316,21 +316,12 @@ def evaluate_credits_notices(
             active.discard(CREDITS_USAGE_KEY)
         if target_band is not None:
             # Belt-and-suspenders: a producer could set subscription_limit_micros
-            # without subscription_limit_usd. Render "$?" rather than "$None".
+            # without subscription_limit_usd. Render "$? cap" rather than "$None cap".
             _cap_usd = state.subscription_limit_usd or "?"
             _level = current_band[1]  # type: ignore[index]  (current_band set when target_band set)
-            # Report absolute dollars used, not a bare "N% used": the percentage is
-            # only meaningful against a Nous subscription cap (no cap → never fires),
-            # so dollars are clearer and don't imply a universal %. Used = cap −
-            # remaining (micros, money-safe), clamped to [0, cap]. Re-emits on band
-            # change (50 → 75 → 90), not every turn — a snapshot, not a live ticker.
-            _lim = state.subscription_limit_micros or 0
-            _used_micros = max(0, min(_lim, _lim - state.subscription_micros))
-            _used_usd = f"{_used_micros / 1_000_000:.2f}" if _lim else "?"
-            _glyph = "⚠" if _level == "warn" else "•"
             to_show.append(
                 AgentNotice(
-                    text=f"{_glyph} You've used ${_used_usd} of your ${_cap_usd} cap",
+                    text=f"{'⚠' if _level == 'warn' else '•'} Credits {target_band}% used · ${_cap_usd} cap",
                     level=_level,
                     kind=CREDITS_NOTICE_KIND,
                     key=CREDITS_USAGE_KEY,
@@ -364,7 +355,7 @@ def evaluate_credits_notices(
     if show_depleted and "credits.depleted" not in active:
         to_show.append(
             AgentNotice(
-                text="✕ Credit access paused · run /topup to top up",
+                text="✕ Credit access paused · run /credits to top up",
                 level="error",
                 kind=CREDITS_NOTICE_KIND,
                 key="credits.depleted",

agent/curator.py

@@ -45,26 +45,12 @@ def _strip_aux_credential(value: Any) -> Optional[str]:
 
 
 class _ReviewRuntimeBinding(NamedTuple):
-    """Provider/model for the curator review fork plus per-slot overrides."""
+    """Provider/model for the curator review fork plus optional per-slot overrides."""
 
     provider: str
     model: str
     explicit_api_key: Optional[str]
     explicit_base_url: Optional[str]
-    request_overrides: Dict[str, Any]
-
-
-def _merge_request_overrides(
-    runtime_overrides: Any,
-    slot_extra_body: Any,
-) -> Dict[str, Any]:
-    """Merge resolver metadata with task-local request body fields."""
-    merged = dict(runtime_overrides or {})
-    if isinstance(slot_extra_body, dict) and slot_extra_body:
-        extra_body = dict(merged.get("extra_body") or {})
-        extra_body.update(slot_extra_body)
-        merged["extra_body"] = extra_body
-    return merged
 
 
 DEFAULT_INTERVAL_HOURS = 24 * 7  # 7 days
@@ -287,21 +273,6 @@ def should_run_now(now: Optional[datetime] = None) -> bool:
 # Automatic state transitions (pure function, no LLM)
 # ---------------------------------------------------------------------------
 
-def _cron_referenced_skills() -> Set[str]:
-    """Skill names referenced by any cron job (incl. paused/disabled).
-
-    Best-effort: a cron-module import error or corrupt jobs store must never
-    break the curator, so any failure yields an empty set (no protection,
-    but no crash).
-    """
-    try:
-        from cron.jobs import referenced_skill_names as _refs
-        return _refs()
-    except Exception as e:
-        logger.debug("Curator could not read cron skill references: %s", e, exc_info=True)
-        return set()
-
-
 def apply_automatic_transitions(now: Optional[datetime] = None) -> Dict[str, int]:
     """Walk every curator-managed skill and move active/stale/archived based on
     the latest real activity timestamp. Pinned skills are never touched.
@@ -321,8 +292,6 @@ def apply_automatic_transitions(now: Optional[datetime] = None) -> Dict[str, int
     stale_cutoff = now - timedelta(days=get_stale_after_days())
     archive_cutoff = now - timedelta(days=get_archive_after_days())
 
-    cron_referenced = _cron_referenced_skills()
-
     counts = {"marked_stale": 0, "archived": 0, "reactivated": 0, "checked": 0, "seeded": 0}
 
     for row in _u.agent_created_report():
@@ -331,15 +300,6 @@ def apply_automatic_transitions(now: Optional[datetime] = None) -> Dict[str, int
         if row.get("pinned"):
             continue
 
-        # A skill referenced by any cron job (incl. paused/disabled) is in
-        # use by definition — resuming or the next fire must find it. The
-        # scheduler only bumps usage when a job actually fires, so jobs that
-        # fire less often than archive_after_days, paused jobs, and far-future
-        # one-shots would otherwise have their skills aged out from under
-        # them. Treat referenced skills like pinned: never auto-transition.
-        if name in cron_referenced:
-            continue
-
         # First sight of a curation-eligible skill with no persisted record
         # (e.g. a newly-eligible built-in): anchor its clock to now and defer.
         if not row.get("_persisted", True):
@@ -356,18 +316,6 @@ def apply_automatic_transitions(now: Optional[datetime] = None) -> Dict[str, int
 
         current = row.get("state", _u.STATE_ACTIVE)
 
-        # Never-used skills (use_count == 0) get a grace floor: don't archive
-        # one until it is at least stale_after_days old. A use=0 skill is
-        # absence of evidence, not evidence of staleness — a skill created
-        # recently may simply not have had its trigger come up yet.
-        never_used = int(row.get("use_count", 0) or 0) == 0
-        if never_used and anchor > stale_cutoff:
-            # Younger than the stale window — leave it alone entirely.
-            if current == _u.STATE_STALE:
-                _u.set_state(name, _u.STATE_ACTIVE)
-                counts["reactivated"] += 1
-            continue
-
         if anchor <= archive_cutoff and current != _u.STATE_ARCHIVED:
             ok, _msg = _u.archive_skill(name)
             if ok:
@@ -422,19 +370,15 @@ CURATOR_REVIEW_PROMPT = (
     "INSTRUCTIONS AND EXPERIENTIAL KNOWLEDGE. A collection of hundreds of "
     "narrow skills where each one captures one session's specific bug is "
     "a FAILURE of the library — not a feature. An agent searching skills "
-    "matches on descriptions, not on exact names (note: long descriptions "
-    "are truncated to 57 chars in the system prompt skill index — keep the "
-    "trigger class in that window). One broad umbrella "
+    "matches on descriptions, not on exact names; one broad umbrella "
     "skill with labeled subsections beats five narrow siblings for "
     "discoverability, not the other way around.\n\n"
     "The right target shape is CLASS-LEVEL skills with rich SKILL.md "
     "bodies + `references/`, `templates/`, and `scripts/` subfiles for "
     "session-specific detail — not one-session-one-skill micro-entries.\n\n"
     "Hard rules — do not violate:\n"
-    "1. DO NOT touch bundled, hub-installed, or external-dir skills "
-    "(`skills.external_dirs`). The candidate list below is already filtered "
-    "to local curator-managed skills only; external skills are externally "
-    "owned and read-only to this background curator.\n"
+    "1. DO NOT touch bundled or hub-installed skills. The candidate list "
+    "below is already filtered to agent-created skills only.\n"
     "2. DO NOT delete any skill. Archiving (moving the skill's directory "
     "into ~/.hermes/skills/.archive/) is the maximum destructive action. "
     "Archives are recoverable; deletion is not.\n"
@@ -444,19 +388,10 @@ CURATOR_REVIEW_PROMPT = (
     "back load-bearing UX (slash-command entry points referenced in docs and "
     "tips) and are filtered out of the candidate list below — never resurrect "
     "one as an archive or absorb target.\n"
-    "3c. DO NOT archive or prune any skill marked `cron=yes` in the candidate "
-    "list. A cron job depends on it and will fail to load it on its next "
-    "run. You MAY still consolidate it into an umbrella — but only because "
-    "the curator rewrites cron job skill references to follow consolidations; "
-    "never simply prune it.\n"
     "4. DO NOT use usage counters as a reason to skip consolidation. The "
     "counters are new and often mostly zero. Judge overlap on CONTENT, "
     "not on use_count. 'use=0' is not evidence a skill is valuable; it's "
-    "absence of evidence either way. Corollary: 'use=0' is ALSO not a "
-    "reason to PRUNE a skill. Never archive a never-used skill (use=0) "
-    "unless it is at least 30 days old (check last_activity / created date) "
-    "AND its content is genuinely obsolete or fully absorbed elsewhere — a "
-    "recently-created skill simply may not have had its trigger come up yet.\n"
+    "absence of evidence either way.\n"
     "5. DO NOT reject consolidation on the grounds that 'each skill has "
     "a distinct trigger'. Pairwise distinctness is the wrong bar. The "
     "right bar is: 'would a human maintainer write this as N separate "
@@ -534,9 +469,8 @@ CURATOR_REVIEW_PROMPT = (
     "skill, or `absorbed_into=\"\"` when you're truly pruning with no "
     "forwarding target. This drives cron-job skill-reference migration — "
     "guessing from your YAML summary after the fact is fragile.\n"
-    "  - terminal                       — move LOCAL candidate content into "
-    "a support subfile when package integrity requires it; never mv, cp, rm, "
-    "patch, or rewrite bundled, hub-installed, or external-dir skills\n\n"
+    "  - terminal                       — mv a sibling into the archive "
+    "OR move its content into a support subfile\n\n"
     "'keep' is a legitimate decision ONLY when the skill is already a "
     "class-level umbrella and none of the proposed merges would improve "
     "discoverability. 'This is narrow but distinct from its siblings' "
@@ -1476,14 +1410,12 @@ def _render_candidate_list() -> str:
     rows = skill_usage.agent_created_report()
     if not rows:
         return "No agent-created skills to review."
-    cron_referenced = _cron_referenced_skills()
     lines = [f"Agent-created skills ({len(rows)}):\n"]
     for r in rows:
         lines.append(
             f"- {r['name']}  "
             f"state={r['state']}  "
             f"pinned={'yes' if r.get('pinned') else 'no'}  "
-            f"cron={'yes' if r['name'] in cron_referenced else 'no'}  "
             f"activity={r.get('activity_count', 0)}  "
             f"use={r.get('use_count', 0)}  "
             f"view={r.get('view_count', 0)}  "
@@ -1780,7 +1712,6 @@ def _resolve_review_runtime(cfg: Dict[str, Any]) -> _ReviewRuntimeBinding:
             _task_model,
             _strip_aux_credential(_cur_task.get("api_key")),
             _strip_aux_credential(_cur_task.get("base_url")),
-            _merge_request_overrides({}, _cur_task.get("extra_body")),
         )
 
     # 2. Legacy curator.auxiliary.{provider,model} (deprecated, pre-unification)
@@ -1798,11 +1729,10 @@ def _resolve_review_runtime(cfg: Dict[str, Any]) -> _ReviewRuntimeBinding:
             str(_legacy_model),
             _strip_aux_credential(_legacy.get("api_key")),
             _strip_aux_credential(_legacy.get("base_url")),
-            _merge_request_overrides({}, _legacy.get("extra_body")),
         )
 
     # 3. Fall through to the main chat model
-    return _ReviewRuntimeBinding(_main_provider, _main_model, None, None, {})
+    return _ReviewRuntimeBinding(_main_provider, _main_model, None, None)
 
 
 def _resolve_review_model(cfg: Dict[str, Any]) -> tuple[str, str]:
@@ -1868,11 +1798,6 @@ def _run_llm_review(prompt: str) -> Dict[str, Any]:
     _base_url = None
     _api_mode = None
     _resolved_provider = None
-    _credential_pool = None
-    _request_overrides: Dict[str, Any] = {}
-    _max_tokens = None
-    _acp_command = None
-    _acp_args = None
     _model_name = ""
     try:
         from hermes_cli.config import load_config
@@ -1890,16 +1815,6 @@ def _run_llm_review(prompt: str) -> Dict[str, Any]:
         _base_url = _rp.get("base_url")
         _api_mode = _rp.get("api_mode")
         _resolved_provider = _rp.get("provider") or _provider
-        _credential_pool = _rp.get("credential_pool")
-        _request_overrides = _merge_request_overrides(
-            _rp.get("request_overrides"),
-            _binding.request_overrides.get("extra_body"),
-        )
-        _max_tokens = _rp.get("max_output_tokens")
-        _acp_command = _rp.get("command")
-        _acp_args = list(_rp.get("args") or [])
-        if isinstance(_rp.get("model"), str) and _rp["model"].strip():
-            _model_name = _rp["model"].strip()
     except Exception as e:
         logger.debug("Curator provider resolution failed: %s", e, exc_info=True)
 
@@ -1908,21 +1823,12 @@ def _run_llm_review(prompt: str) -> Dict[str, Any]:
 
     review_agent = None
     try:
-        _agent_kwargs: Dict[str, Any] = {}
-        if isinstance(_max_tokens, int):
-            _agent_kwargs["max_tokens"] = _max_tokens
-        if isinstance(_acp_command, str) and _acp_command:
-            _agent_kwargs["acp_command"] = _acp_command
-            _agent_kwargs["acp_args"] = _acp_args or []
         review_agent = AIAgent(
             model=_model_name,
             provider=_resolved_provider,
             api_key=_api_key,
             base_url=_base_url,
             api_mode=_api_mode,
-            credential_pool=_credential_pool,
-            request_overrides=_request_overrides,
-            **_agent_kwargs,
             # Umbrella-building over a large skill collection is worth a
             # high iteration ceiling — the pass typically takes 50-100
             # API calls against hundreds of candidate skills. The
@@ -1937,14 +1843,6 @@ def _run_llm_review(prompt: str) -> Dict[str, Any]:
         # Disable recursive nudges — the curator must never spawn its own review.
         review_agent._memory_nudge_interval = 0
         review_agent._skill_nudge_interval = 0
-        # Tag this fork as autonomous background curation so skill_manage's
-        # background-review write guard fires. Without this the fork inherits
-        # the default "assistant_tool" origin, is_background_review() is False,
-        # and the external/bundled/hub-installed skill_manage guards never
-        # trigger during the curation pass they exist to protect against.
-        # turn_context.py binds this onto the write-origin ContextVar at turn
-        # start (see agent/turn_context.py).
-        review_agent._memory_write_origin = "background_review"
 
         # Redirect the forked agent's stdout/stderr to /dev/null while it
         # runs so its tool-call chatter doesn't pollute the foreground

agent/curator_backup.py

@@ -98,12 +98,7 @@ def _backup_cron_jobs_into(dest: Path) -> Dict[str, Any]:
         info["reason"] = "no cron/jobs.json present"
         return info
     try:
-        # utf-8-sig: same dialect as cron/jobs.load_jobs — a UTF-8 BOM left
-        # by Windows editors otherwise survives decoding as U+FEFF, breaks
-        # json.loads below, and misreports jobs_count as 0 with a spurious
-        # parse warning. The BOM-less text is also what gets written to the
-        # backup, so a later rollback restores a loadable file.
-        raw = src.read_text(encoding="utf-8-sig")
+        raw = src.read_text(encoding="utf-8")
     except OSError as e:
         logger.debug("Failed to read cron/jobs.json for backup: %s", e)
         info["reason"] = f"read error: {e}"
@@ -561,7 +556,7 @@ def rollback(backup_id: Optional[str] = None) -> Tuple[bool, str, Optional[Path]
     if target is None:
         return (
             False,
-            "no matching backup found"
+            f"no matching backup found"
             + (f" for id '{backup_id}'" if backup_id else "")
             + " (use `hermes curator rollback --list` to see available snapshots)",
             None,

agent/delegation_context.py

@@ -1,85 +0,0 @@
-"""Context-local state for delegate_task child execution.
-
-The parent Hermes process may itself be a Kanban dispatcher worker with
-HERMES_KANBAN_* variables in process env. delegate_task children run inside the
-same Python process, but they are not dispatcher-owned Kanban workers. This
-module lets code paths that resolve tool schemas or spawn subprocesses fail
-closed for delegated children without mutating global os.environ for the parent.
-"""
-from __future__ import annotations
-
-from contextlib import contextmanager
-from contextvars import ContextVar
-from typing import Iterator, Mapping, MutableMapping
-
-_DELEGATED_CHILD_CONTEXT: ContextVar[bool] = ContextVar(
-    "hermes_delegated_child_context",
-    default=False,
-)
-
-DELEGATED_CHILD_ENV_MARKER = "HERMES_DELEGATED_CHILD_CONTEXT"
-
-KANBAN_ENV_KEYS: tuple[str, ...] = (
-    "HERMES_KANBAN_TASK",
-    "HERMES_KANBAN_RUN_ID",
-    "HERMES_KANBAN_WORKSPACE",
-    "HERMES_KANBAN_WORKSPACES_ROOT",
-    "HERMES_KANBAN_CLAIM_LOCK",
-    "HERMES_KANBAN_BOARD",
-    "HERMES_KANBAN_DB",
-)
-
-
-@contextmanager
-def delegated_child_context() -> Iterator[None]:
-    """Mark the current execution context as a delegate_task child."""
-    token = _DELEGATED_CHILD_CONTEXT.set(True)
-    try:
-        yield
-    finally:
-        _DELEGATED_CHILD_CONTEXT.reset(token)
-
-
-def is_delegated_child_context() -> bool:
-    """Return True while code is running for a delegate_task child."""
-    return bool(_DELEGATED_CHILD_CONTEXT.get())
-
-
-def is_delegated_child_process_context() -> bool:
-    """Return True in this process or a subprocess spawned by a child."""
-    import os
-
-    return bool(_DELEGATED_CHILD_CONTEXT.get()) or bool(
-        os.environ.get(DELEGATED_CHILD_ENV_MARKER)
-    )
-
-
-def scrub_kanban_env(env: Mapping[str, str] | MutableMapping[str, str]) -> dict[str, str]:
-    """Return *env* with dispatcher-only Kanban variables removed."""
-    cleaned = dict(env)
-    for key in KANBAN_ENV_KEYS:
-        cleaned.pop(key, None)
-    cleaned[DELEGATED_CHILD_ENV_MARKER] = "1"
-    return cleaned
-
-
-def delegated_child_subprocess_env(
-    env: Mapping[str, str] | MutableMapping[str, str] | None = None,
-) -> dict[str, str] | None:
-    """Return an env override only when delegated-child lineage must cross fork.
-
-    Most subprocess call sites historically used ``env=None`` to inherit the
-    process environment.  In a ``delegate_task`` child, inheriting as-is leaks
-    parent dispatcher ``HERMES_KANBAN_*`` vars while losing the ContextVar in
-    the new process.  This helper preserves normal ``env=None`` semantics for
-    non-delegated calls, and only materializes a scrubbed env when the lineage
-    marker must be propagated across a child-process boundary.
-    """
-    if not is_delegated_child_process_context():
-        return None if env is None else dict(env)
-
-    if env is None:
-        import os
-
-        env = os.environ
-    return scrub_kanban_env(env)

agent/display.py

@@ -6,7 +6,6 @@ Used by AIAgent._execute_tool_calls for CLI feedback.
 
 import logging
 import os
-import re
 import sys
 import threading
 import time
@@ -16,7 +15,6 @@ from pathlib import Path
 from typing import Any
 
 from utils import safe_json_loads
-from agent.redact import redact_sensitive_text
 from agent.tool_result_classification import file_mutation_result_landed
 
 # ANSI escape codes for coloring tool failure indicators
@@ -27,14 +25,6 @@ logger = logging.getLogger(__name__)
 
 _ANSI_RESET = "\033[0m"
 
-
-def _display_url(value: Any) -> str:
-    """Extract a display-only URL without assuming model argument types."""
-    if isinstance(value, dict):
-        value = value.get("url") or value.get("href")
-    return value.strip() if isinstance(value, str) else ""
-
-
 # Diff colors — resolved lazily from the skin engine so they adapt
 # to light/dark themes.  Falls back to sensible defaults on import
 # failure.  We cache after first resolution for performance.
@@ -187,223 +177,6 @@ def _truncate_preview(text: str, max_len: int | None) -> str:
     return text
 
 
-_SHELL_SILENT_HEADS = {"cd", "pushd", "popd", "export", "set", "unset", "source", ".", "true", "false", ":"}
-_SHELL_PIPE_TAIL_HEADS = {"head", "tail", "wc", "sort", "uniq"}
-
-
-def _shell_basename(head: str) -> str:
-    return head.rsplit("/", 1)[-1] if head else ""
-
-
-def _split_shell_words(segment: str) -> list[str]:
-    words: list[str] = []
-    buf: list[str] = []
-    quote: str | None = None
-
-    for i, ch in enumerate(segment):
-        if quote:
-            buf.append(ch)
-            if ch == quote and (i == 0 or segment[i - 1] != "\\"):
-                quote = None
-            continue
-
-        if ch in {"'", '"'}:
-            quote = ch
-            buf.append(ch)
-            continue
-
-        if ch.isspace():
-            if buf:
-                words.append("".join(buf))
-                buf = []
-            continue
-
-        buf.append(ch)
-
-    if buf:
-        words.append("".join(buf))
-
-    return words
-
-
-def _strip_shell_pipe_tail(segment: str) -> str:
-    words = _split_shell_words(segment)
-    out: list[str] = []
-
-    for i, word in enumerate(words):
-        if word == "|" and _shell_basename(words[i + 1] if i + 1 < len(words) else "") in _SHELL_PIPE_TAIL_HEADS:
-            break
-        out.append(word)
-
-    return " ".join(out).strip()
-
-
-def _split_shell_compound(command: str) -> list[str]:
-    segments: list[str] = []
-    buf: list[str] = []
-    quote: str | None = None
-    i = 0
-
-    while i < len(command):
-        ch = command[i]
-
-        if quote:
-            buf.append(ch)
-            if ch == quote and (i == 0 or command[i - 1] != "\\"):
-                quote = None
-            i += 1
-            continue
-
-        if ch in {"'", '"'}:
-            quote = ch
-            buf.append(ch)
-            i += 1
-            continue
-
-        op_len = 2 if command.startswith("&&", i) or command.startswith("||", i) else 1 if ch in {";", "\n"} else 0
-        if op_len:
-            segment = _strip_shell_pipe_tail("".join(buf).strip())
-            if segment:
-                segments.append(segment)
-            buf = []
-            i += op_len
-            continue
-
-        buf.append(ch)
-        i += 1
-
-    segment = _strip_shell_pipe_tail("".join(buf).strip())
-    if segment:
-        segments.append(segment)
-
-    return segments
-
-
-def _shell_head_word(segment: str) -> str:
-    words = _split_shell_words(segment)
-    index = 0
-    while index < len(words) and re.match(r"^[A-Za-z_]\w*=", words[index]):
-        index += 1
-    return _shell_basename(words[index] if index < len(words) else "")
-
-
-def _clean_shell_segment(segment: str) -> str:
-    words = _split_shell_words(segment)
-    out: list[str] = []
-    i = 0
-    while i < len(words):
-        word = words[i]
-        if re.match(r"^\d*(?:>>?|<)$", word):
-            i += 2
-            continue
-        if re.match(r"^\d*(?:>&|<&)\d+$", word) or re.match(r"^\d*>&\d+$", word):
-            i += 1
-            continue
-        out.append(word)
-        i += 1
-    return " ".join(out).strip()
-
-
-def _is_shell_boundary_echo(segment: str) -> bool:
-    words = _split_shell_words(segment)
-    if _shell_basename(words[0] if words else "") != "echo":
-        return False
-    rest = " ".join(words[1:])
-    return bool(re.search(r"-{2,}|_exit=|(?:^|\s|=)\$[?{]|PIPESTATUS", rest))
-
-
-def summarize_shell_command(command: str) -> str:
-    """Compact shell wrapper/plumbing for display while preserving raw command elsewhere."""
-    original = _oneline(command)
-    if not original:
-        return ""
-
-    segments = _split_shell_compound(original)
-    if len(segments) <= 1:
-        return _clean_shell_segment(segments[0] if segments else original) or original
-
-    core: list[str] = []
-    for segment in segments:
-        cleaned = _clean_shell_segment(segment)
-        head = _shell_head_word(cleaned)
-        if cleaned and head not in _SHELL_SILENT_HEADS and not _is_shell_boundary_echo(cleaned):
-            core.append(cleaned)
-
-    if not core:
-        return original
-    if len(core) == 1:
-        return core[0]
-
-    count = len(core) - 1
-    return f"{core[0]} + {count} {'command' if count == 1 else 'commands'}"
-
-
-def _read_file_line_label(args: dict) -> str:
-    offset = args.get("offset")
-    limit = args.get("limit")
-    if not isinstance(offset, int) or offset <= 0:
-        return ""
-    if not isinstance(limit, int) or limit <= 1:
-        return f"L{offset}"
-    return f"L{offset}-{offset + limit - 1}"
-
-
-def redact_browser_typed_text_for_display(value: Any, typed_text: Any) -> Any:
-    """Apply secret redaction to browser_type text in display-facing payloads.
-
-    Backends sometimes echo the attempted input in error strings or fallback
-    metadata.  When the raw typed value contains a recognizable secret (API
-    key, token, JWT, etc.) the redacted form differs from the raw value, so we
-    replace every occurrence of the raw value with its redacted form before a
-    browser_type result reaches logs, callbacks, the model, or chat history.
-
-    Normal typed text (search queries, addresses, form fields) matches no
-    secret pattern, so it passes through unchanged and stays readable.
-
-    Redaction is forced here regardless of the global ``security.redact_secrets``
-    preference: a typed credential leaking into chat history is a security
-    boundary, not mere log hygiene.
-    """
-    if typed_text is None:
-        return value
-    needle = str(typed_text)
-    if needle == "":
-        return value
-    redacted = redact_sensitive_text(needle, force=True)
-    if redacted == needle:
-        # Nothing secret-looking in the typed text; leave payload untouched.
-        return value
-    if isinstance(value, str):
-        return value.replace(needle, redacted)
-    if isinstance(value, dict):
-        return {
-            key: redact_browser_typed_text_for_display(item, typed_text)
-            for key, item in value.items()
-        }
-    if isinstance(value, list):
-        return [redact_browser_typed_text_for_display(item, typed_text) for item in value]
-    if isinstance(value, tuple):
-        return tuple(redact_browser_typed_text_for_display(item, typed_text) for item in value)
-    return value
-
-
-def redact_tool_args_for_display(tool_name: str, args: dict | None) -> dict | None:
-    """Return a copy of tool args safe for logs/progress UI.
-
-    For ``browser_type`` the ``text`` argument is run through the same
-    secret-pattern redactor used for logs.  Recognizable credentials (API
-    keys, tokens) are masked before the value reaches tool progress
-    notifications; normal typed text is left intact for debuggability.
-    """
-    if not isinstance(args, dict):
-        return args
-    if tool_name == "browser_type" and isinstance(args.get("text"), str):
-        safe_args = dict(args)
-        safe_args["text"] = redact_sensitive_text(args["text"], force=True)
-        return safe_args
-    return args
-
-
 def _delegate_task_goal_parts(tasks: Any, *, per_goal_len: int) -> tuple[int, list[str]]:
     if not isinstance(tasks, list):
         return 0, []
@@ -427,14 +200,13 @@ def build_tool_preview(tool_name: str, args: dict, max_len: int | None = None) -
         max_len = _tool_preview_max_len
     if not args:
         return None
-    args = redact_tool_args_for_display(tool_name, args) or args
     primary_args = {
         "terminal": "command", "web_search": "query", "web_extract": "urls",
         "read_file": "path", "write_file": "path", "patch": "path",
         "search_files": "pattern", "browser_navigate": "url",
         "browser_click": "ref", "browser_type": "text",
         "image_generate": "prompt", "text_to_speech": "text",
-        "vision_analyze": "question",
+        "vision_analyze": "question", "mixture_of_agents": "user_prompt",
         "skill_view": "name", "skills_list": "category",
         "cronjob": "action",
         "execute_code": "code", "delegate_task": "goal",
@@ -462,14 +234,13 @@ def build_tool_preview(tool_name: str, args: dict, max_len: int | None = None) -
         sid = args.get("session_id", "")
         data = args.get("data", "")
         timeout_val = args.get("timeout")
-        parts = [str(action) if action else ""]
+        parts = [action]
         if sid:
-            parts.append(str(sid)[:16])
+            parts.append(sid[:16])
         if data:
-            parts.append(f'"{_oneline(str(data)[:20])}"')
+            parts.append(f'"{_oneline(data[:20])}"')
         if timeout_val and action == "wait":
             parts.append(f"{timeout_val}s")
-        parts = [p for p in parts if p]
         return " ".join(parts) if parts else None
 
     if tool_name == "todo":
@@ -482,23 +253,6 @@ def build_tool_preview(tool_name: str, args: dict, max_len: int | None = None) -
         else:
             return f"planning {len(todos_arg)} task(s)"
 
-    if tool_name in {"terminal", "execute_code"}:
-        key = "code" if tool_name == "execute_code" else "command"
-        command = args.get(key)
-        if command is None:
-            return None
-        preview = summarize_shell_command(str(command))
-        return _truncate_preview(preview, max_len) if preview else None
-
-    if tool_name == "read_file":
-        path = args.get("path") or args.get("file") or args.get("filepath")
-        if path is None:
-            return None
-        label = Path(str(path).replace("\\", "/")).name or str(path)
-        line_label = _read_file_line_label(args)
-        preview = f"{label} {line_label}".strip()
-        return _truncate_preview(preview, max_len) if preview else None
-
     if tool_name == "session_search":
         query = _oneline(args.get("query", ""))
         return f"recall: \"{query[:25]}{'...' if len(query) > 25 else ''}\""
@@ -524,16 +278,6 @@ def build_tool_preview(tool_name: str, args: dict, max_len: int | None = None) -
             msg = msg[:17] + "..."
         return f"to {target}: \"{msg}\""
 
-    if tool_name == "skill_view":
-        name = _oneline(str(args.get("name") or ""))
-        file_path = args.get("file_path")
-        if file_path:
-            file_path = _oneline(str(file_path))
-            preview = f"{name} → {file_path}" if name else file_path
-        else:
-            preview = name
-        return _truncate_preview(preview, max_len) if preview else None
-
     key = primary_args.get(tool_name)
     if not key:
         for fallback_key in ("query", "text", "command", "path", "name", "prompt", "code", "goal"):
@@ -556,168 +300,6 @@ def build_tool_preview(tool_name: str, args: dict, max_len: int | None = None) -
     return preview
 
 
-# =========================================================================
-# Friendly tool labels (human-phrased verbs for built-in tools)
-#
-# Turns "web_search <query>" into "Searching the web for <query>" — the
-# ChatGPT-style "Searching…/Reading…" surface.  Curated and built-in only:
-# we know each core tool's semantics, so the verb is fixed, not computed.
-# Custom/plugin/MCP tools have no entry and fall back to the raw preview.
-# =========================================================================
-
-# Each entry maps a built-in tool name to its present-participle verb phrase.
-# A trailing space-then-preview is appended by build_tool_label() when the
-# tool's argument preview is available (e.g. "Reading docs/api.md").
-_TOOL_VERBS: dict[str, str] = {
-    "web_search": "Searching the web",
-    "web_extract": "Reading",
-    "browser_navigate": "Browsing",
-    "browser_click": "Clicking",
-    "browser_type": "Typing",
-    "read_file": "Reading",
-    "write_file": "Writing",
-    "patch": "Editing",
-    "search_files": "Searching files",
-    "terminal": "Running",
-    "execute_code": "Running code",
-    "image_generate": "Generating image",
-    "video_generate": "Generating video",
-    "text_to_speech": "Generating speech",
-    "vision_analyze": "Looking at the image",
-    "session_search": "Searching past sessions",
-    "skill_view": "Reading skill",
-    "skills_list": "Listing skills",
-    "skill_manage": "Updating skill",
-    "delegate_task": "Delegating",
-    "cronjob": "Scheduling",
-    "clarify": "Asking",
-    "memory": "Updating memory",
-    "todo": "Updating tasks",
-}
-
-# Verbs that read better without the raw argument preview appended.
-_TOOL_VERBS_NO_PREVIEW: frozenset[str] = frozenset({
-    "skills_list",
-    "session_search",
-})
-
-# Verbs that take a "for" connector before the preview (search-style phrasing):
-# "Searching the web for <query>" reads better than "Searching the web <query>".
-_TOOL_VERBS_FOR_CONNECTOR: frozenset[str] = frozenset({
-    "web_search",
-    "search_files",
-})
-
-_friendly_tool_labels: bool = True
-
-
-def set_friendly_tool_labels(enabled: bool) -> None:
-    """Toggle friendly human-phrased tool labels (display.friendly_tool_labels)."""
-    global _friendly_tool_labels
-    _friendly_tool_labels = bool(enabled)
-
-
-def get_friendly_tool_labels() -> bool:
-    """Return whether friendly tool labels are enabled."""
-    return _friendly_tool_labels
-
-
-def get_tool_verb(tool_name: str) -> str | None:
-    """Return the friendly verb for a built-in tool, or None.
-
-    Returns None when friendly labels are disabled or the tool has no curated
-    verb (custom/plugin/MCP tools).  Callers that already hold a computed
-    argument preview can compose ``f"{verb} {preview}"`` themselves; use
-    :func:`tool_verb_connector` to pick the right joiner.
-    """
-    if not _friendly_tool_labels:
-        return None
-    return _TOOL_VERBS.get(tool_name)
-
-
-def tool_verb_connector(tool_name: str) -> str:
-    """Return the connector between a verb and its preview (" for " or " ")."""
-    return " for " if tool_name in _TOOL_VERBS_FOR_CONNECTOR else " "
-
-
-def verb_drops_preview(tool_name: str) -> bool:
-    """Whether the verb should render alone, without the argument preview."""
-    return tool_name in _TOOL_VERBS_NO_PREVIEW
-
-
-def build_status_phrase(tool_name: str, args: dict | None, max_len: int = 49) -> str | None:
-    """Build a short present-tense status phrase for platform status surfaces.
-
-    Used by text-rendering "typing" indicators (Slack's
-    ``assistant.threads.setStatus`` line) to show what the agent is doing
-    right now: ``is running scripts/run_tests.sh…`` instead of a static
-    ``is thinking...``.  The phrase is phrased to follow the bot's display
-    name ("Hermes is running …"), so it starts lowercase with "is".
-
-    Pass ``args=None`` for a verb-only phrase (``is running…``) — used when
-    ``display.live_status`` is ``verb`` to keep argument previews out of
-    shared channels.
-
-    Returns None for the ``_thinking`` pseudo-tool and when friendly labels
-    are disabled (callers fall back to their static default).  ``max_len``
-    caps the total phrase length; Slack truncates its status line around 50
-    characters, so the default stays just under that.
-    """
-    if not tool_name or tool_name == "_thinking":
-        return None
-    if not _friendly_tool_labels:
-        return None
-
-    verb = _TOOL_VERBS.get(tool_name)
-    if verb:
-        head = f"is {verb[0].lower()}{verb[1:]}"
-    else:
-        # Custom / plugin / MCP tools: generic but still informative.
-        head = f"is using {tool_name}"
-
-    phrase = head
-    if args and verb and tool_name not in _TOOL_VERBS_NO_PREVIEW:
-        preview = build_tool_preview(tool_name, args, max_len=None)
-        if preview:
-            # Previews can contain newlines (terminal commands); keep the
-            # status to the first line.
-            preview = preview.splitlines()[0].strip()
-            phrase = f"{head}{tool_verb_connector(tool_name)}{preview}"
-
-    if len(phrase) > max_len - 1:
-        phrase = phrase[: max_len - 2].rstrip() + "…"
-    else:
-        phrase = phrase + "…"
-    return phrase
-
-
-def build_tool_label(tool_name: str, args: dict, max_len: int | None = None) -> str | None:
-    """Build a human-phrased status label for a tool call.
-
-    For built-in tools with a known verb (``web_search`` -> "Searching the
-    web for ..."), returns the verb optionally followed by the argument
-    preview.  For everything else (custom/plugin/MCP tools, or when friendly
-    labels are disabled) returns the raw preview, so callers can use this as a
-    drop-in replacement for :func:`build_tool_preview`.
-    """
-    if not _friendly_tool_labels:
-        return build_tool_preview(tool_name, args, max_len=max_len)
-
-    verb = _TOOL_VERBS.get(tool_name)
-    if not verb:
-        return build_tool_preview(tool_name, args, max_len=max_len)
-
-    if tool_name in _TOOL_VERBS_NO_PREVIEW:
-        return verb
-
-    preview = build_tool_preview(tool_name, args, max_len=max_len)
-    if not preview:
-        return verb
-    if tool_name in _TOOL_VERBS_FOR_CONNECTOR:
-        return f"{verb} for {preview}"
-    return f"{verb} {preview}"
-
-
 # =========================================================================
 # Inline diff previews for write actions
 # =========================================================================
@@ -1314,7 +896,7 @@ def _detect_tool_failure(tool_name: str, result: str | None) -> tuple[bool, str]
     return False, ""
 
 
-def _get_cute_tool_message(
+def get_cute_tool_message(
     tool_name: str, args: dict, duration: float, result: str | None = None,
 ) -> str:
     """Generate a formatted tool completion line for CLI quiet mode.
@@ -1324,7 +906,6 @@ def _get_cute_tool_message(
     When *result* is provided the line is checked for failure indicators.
     Failed tool calls get a red prefix and an informational suffix.
     """
-    args = redact_tool_args_for_display(tool_name, args) or args
     dur = f"{duration:.1f}s"
     is_failure, failure_suffix = _detect_tool_failure(tool_name, result)
     skin_prefix = get_skin_tool_prefix()
@@ -1356,15 +937,13 @@ def _get_cute_tool_message(
     if tool_name == "web_extract":
         urls = args.get("urls", [])
         if urls:
-            url = _display_url(urls[0] if isinstance(urls, list) else urls)
-            if not url:
-                return _wrap(f"┊ 📄 fetch     pages  {dur}")
+            url = urls[0] if isinstance(urls, list) else str(urls)
             domain = url.replace("https://", "").replace("http://", "").split("/")[0]
-            extra = f" +{len(urls)-1}" if isinstance(urls, list) and len(urls) > 1 else ""
+            extra = f" +{len(urls)-1}" if len(urls) > 1 else ""
             return _wrap(f"┊ 📄 fetch     {_trunc(domain, 35)}{extra}  {dur}")
         return _wrap(f"┊ 📄 fetch     pages  {dur}")
     if tool_name == "terminal":
-        return _wrap(f"┊ 💻 $         {_trunc(build_tool_preview(tool_name, args) or args.get('command', ''), 42)}  {dur}")
+        return _wrap(f"┊ 💻 $         {_trunc(args.get('command', ''), 42)}  {dur}")
     if tool_name == "process":
         action = args.get("action", "?")
         sid = args.get("session_id", "")[:12]
@@ -1372,7 +951,7 @@ def _get_cute_tool_message(
                   "wait": f"wait {sid}", "kill": f"kill {sid}", "write": f"write {sid}", "submit": f"submit {sid}"}
         return _wrap(f"┊ ⚙️  proc      {labels.get(action, f'{action} {sid}')}  {dur}")
     if tool_name == "read_file":
-        return _wrap(f"┊ 📖 read      {_trunc(build_tool_preview(tool_name, args) or args.get('path', ''), 42)}  {dur}")
+        return _wrap(f"┊ 📖 read      {_path(args.get('path', ''))}  {dur}")
     if tool_name == "write_file":
         return _wrap(f"┊ ✍️  write     {_path(args.get('path', ''))}  {dur}")
     if tool_name == "patch":
@@ -1451,17 +1030,15 @@ def _get_cute_tool_message(
     if tool_name == "skills_list":
         return _wrap(f"┊ 📚 skills    list {args.get('category', 'all')}  {dur}")
     if tool_name == "skill_view":
-        label = args.get("name", "")
-        file_path = args.get("file_path")
-        if file_path:
-            label = f"{label} → {file_path}" if label else str(file_path)
-        return _wrap(f"┊ 📚 skill     {_trunc(label, 44)}  {dur}")
+        return _wrap(f"┊ 📚 skill     {_trunc(args.get('name', ''), 30)}  {dur}")
     if tool_name == "image_generate":
         return _wrap(f"┊ 🎨 create    {_trunc(args.get('prompt', ''), 35)}  {dur}")
     if tool_name == "text_to_speech":
         return _wrap(f"┊ 🔊 speak     {_trunc(args.get('text', ''), 30)}  {dur}")
     if tool_name == "vision_analyze":
         return _wrap(f"┊ 👁️  vision    {_trunc(args.get('question', ''), 30)}  {dur}")
+    if tool_name == "mixture_of_agents":
+        return _wrap(f"┊ 🧠 reason    {_trunc(args.get('user_prompt', ''), 30)}  {dur}")
     if tool_name == "send_message":
         return _wrap(f"┊ 📨 send      {args.get('target', '?')}: \"{_trunc(args.get('message', ''), 25)}\"  {dur}")
     if tool_name == "cronjob":
@@ -1490,19 +1067,6 @@ def _get_cute_tool_message(
     return _wrap(f"┊ ⚡ {tool_name[:9]:9} {_trunc(preview, 35)}  {dur}")
 
 
-def get_cute_tool_message(
-    tool_name: str, args: dict, duration: float, result: str | None = None,
-) -> str:
-    """Render a completion label without letting cosmetic failures escape."""
-    try:
-        return _get_cute_tool_message(tool_name, args, duration, result=result)
-    except Exception as exc:  # noqa: BLE001 — display must never abort a turn
-        logger.debug("Tool completion label failed for %s: %s", tool_name, exc)
-        safe_name = tool_name[:9] if isinstance(tool_name, str) and tool_name else "tool"
-        safe_duration = f"{duration:.1f}s" if isinstance(duration, (int, float)) else "done"
-        return f"┊ ⚡ {safe_name:9} completed  {safe_duration}"
-
-
 # =========================================================================
 # Honcho session line (one-liner with clickable OSC 8 hyperlink)
 # =========================================================================

agent/error_classifier.py

@@ -31,9 +31,6 @@ class FailoverReason(enum.Enum):
     # Billing / quota
     billing = "billing"                  # 402 or confirmed credit exhaustion — rotate immediately
     rate_limit = "rate_limit"            # 429 or quota-based throttling — backoff then rotate
-    # Upstream model rate-limited (aggregator 429) — fallback to a different
-    # model, NOT credential rotation. The user's key is healthy.
-    upstream_rate_limit = "upstream_rate_limit"
 
     # Server-side
     overloaded = "overloaded"            # 503/529 — provider overloaded, backoff
@@ -41,11 +38,6 @@ class FailoverReason(enum.Enum):
 
     # Transport
     timeout = "timeout"                  # Connection/read timeout — rebuild client + retry
-    # TLS certificate verification failure — deterministic for the host
-    # (TLS-inspecting proxy, missing/expired CA bundle, self-signed cert).
-    # Retrying reproduces the identical handshake failure, so fail fast
-    # with actionable guidance instead of burning retries.
-    ssl_cert_verification = "ssl_cert_verification"
 
     # Context / payload
     context_overflow = "context_overflow"  # Context too large — compress, not failover
@@ -115,7 +107,6 @@ _BILLING_PATTERNS = [
     "exceeded your current quota",
     "account is deactivated",
     "plan does not include",
-    "out of extra usage",  # Anthropic OAuth Pro/Max overage bucket depleted (HTTP 400)
     "out of funds",
     "run out of funds",
     "balance_depleted",
@@ -123,25 +114,6 @@ _BILLING_PATTERNS = [
     "not available on the free tier",
 ]
 
-# xAI's explicit Grok credit-exhaustion code. Keep the HTTP 403 special case
-# provider-scoped: other providers' generic billing codes historically remain
-# auth failures when they arrive as 403.
-_XAI_SPENDING_LIMIT_ERROR_CODE = "personal-team-blocked:spending-limit"
-
-# Structured provider codes that mean the account cannot serve paid traffic
-# until credits/subscription capacity is restored. xAI returns its explicit
-# Grok spending-limit signal as HTTP 403 rather than 402.
-_BILLING_ERROR_CODES = frozenset({
-    "insufficient_quota",
-    "billing_not_active",
-    "payment_required",
-    "insufficient_credits",
-    "no_usable_credits",
-    "balance_depleted",
-    "model_not_supported_on_free_tier",
-    _XAI_SPENDING_LIMIT_ERROR_CODE,
-})
-
 # Patterns that indicate rate limiting (transient, will resolve)
 _RATE_LIMIT_PATTERNS = [
     "rate limit",
@@ -161,31 +133,6 @@ _RATE_LIMIT_PATTERNS = [
     "servicequotaexceededexception",
 ]
 
-# Patterns that indicate provider-side overload, NOT a per-credential rate
-# limit or billing problem.  The credential is valid — the server is just
-# busy — so the correct recovery is "back off and retry the same key", never
-# "rotate the credential" (rotating exhausts the pool while the endpoint is
-# still busy; a single-key user has nothing to rotate to).  Some providers
-# (notably Z.AI / Zhipu) reuse HTTP 429 for server-wide overload, so the 429
-# status path matches the body against this list before falling through to
-# the rate_limit default.  Phrases are kept narrow and overload-flavoured so a
-# normal rate-limit message ("you have been rate-limited") doesn't hit this
-# bucket. (#14038, #15297)
-_OVERLOADED_PATTERNS = [
-    "overloaded",
-    "temporarily overloaded",
-    "service is temporarily overloaded",
-    "service may be temporarily overloaded",
-    "server is overloaded",
-    "server overloaded",
-    "service overloaded",
-    "service is overloaded",
-    "upstream overloaded",
-    "currently overloaded",
-    "at capacity",
-    "over capacity",
-]
-
 # Usage-limit patterns that need disambiguation (could be billing OR rate_limit)
 _USAGE_LIMIT_PATTERNS = [
     "usage limit",
@@ -269,11 +216,6 @@ _CONTEXT_OVERFLOW_PATTERNS = [
     "context window",
     "prompt is too long",
     "prompt exceeds max length",
-    # NOTE: bare "max_tokens" is load-bearing — the output-cap-retry path keys
-    # off it (e.g. "max_tokens: 65536 > context_window: 200000 ..."). Do NOT
-    # remove it. Provider empty-response advisories also contain "very low
-    # max_tokens", but those are intercepted by _EMPTY_PROVIDER_RESPONSE_PATTERNS
-    # BEFORE this list is consulted, so they never mis-route into compression.
     "max_tokens",
     "maximum number of tokens",
     # vLLM / local inference server patterns
@@ -291,8 +233,6 @@ _CONTEXT_OVERFLOW_PATTERNS = [
     # Chinese error messages (some providers return these)
     "超过最大长度",
     "上下文长度",
-    # Z.AI / Zhipu GLM pattern (English form; error code 1210)
-    "tokens in request more than max tokens allowed",
     # AWS Bedrock Converse API error patterns
     "input is too long",
     "max input token",
@@ -310,15 +250,6 @@ _MODEL_NOT_FOUND_PATTERNS = [
     "no such model",
     "unknown model",
     "unsupported model",
-    # OpenRouter returns 404 with this message when none of the candidate
-    # endpoints for the selected model support tool/function calling.
-    # Classifying this as model_not_found triggers fallback to a different
-    # model or provider that does support tools.  Without this entry the
-    # pattern falls through to ``unknown`` with ``retryable=True``, the
-    # retry loop burns all attempts on the same deterministic rejection,
-    # and the error surfaces as a confusing "model not found" message
-    # instead of automatically failing over.  See PR #58446.
-    "no endpoints found that support tool use",
 ]
 
 # Request-validation patterns — the request is malformed and will fail
@@ -399,21 +330,12 @@ _CONTENT_POLICY_BLOCKED_PATTERNS = [
     # echo back; the underscore form is provider-specific enough.
     "content_filter",
     "responsibleaipolicyviolation",
-    # MiniMax output-layer safety filter. The error string is surfaced
-    # verbatim by MiniMax SDK / OpenAI-compatible endpoints, usually in the
-    # form "output new_sensitive (1027)" when the model's *output* (often a
-    # large tool-call argument block) trips the upstream safety filter and
-    # the SSE stream is truncated mid-flight. ``new_sensitive`` is the
-    # filter name and is narrow enough that billing / format / auth error
-    # strings will not collide. See #32421.
-    "new_sensitive",
 ]
 
 # Auth patterns (non-status-code signals)
 _AUTH_PATTERNS = [
     "invalid api key",
     "invalid_api_key",
-    "gateway_auth_failed",
     "authentication",
     "unauthorized",
     "forbidden",
@@ -432,19 +354,6 @@ _THINKING_SIG_PATTERNS = [
 # the exception type is generic (e.g. RuntimeError from a local shim that
 # wraps a subprocess timeout).  Checked before the type-based transport
 # heuristics so custom-provider "timed out" errors don't fall through to
-# Provider empty-response advisories (OpenRouter / nano-gpt / similar).
-# Checked before context-overflow matching because the advisory text often
-# mentions "max_tokens" as a possible cause, which historically sat in
-# _CONTEXT_OVERFLOW_PATTERNS and sent healthy sessions into a compression
-# death spiral ending in "Cannot compress further".
-_EMPTY_PROVIDER_RESPONSE_PATTERNS = [
-    "returned an empty response",
-    "empty response despite retries",
-    "provider returned an empty response",
-    "model returning empty responses",
-    "empty response stream",
-]
-
 # the unknown bucket and get misreported as empty responses.
 _TIMEOUT_MESSAGE_PATTERNS = [
     "timed out",
@@ -492,29 +401,6 @@ _SERVER_DISCONNECT_PATTERNS = [
     "incomplete chunked read",
 ]
 
-# SSL certificate verification failures — deterministic, NOT transient.
-#
-# A failed certificate chain (TLS-inspecting corporate proxy, missing
-# custom CA in the trust store, expired certificate, self-signed cert)
-# fails identically on every retry. Burning the retry budget before
-# surfacing the error hides the actionable fix from the user for minutes.
-# Inspired by Claude Code v2.1.199 (July 2026), which made SSL certificate
-# errors fail immediately with a fix hint instead of retrying.
-#
-# Must be checked BEFORE _SSL_TRANSIENT_PATTERNS — "certificate verify
-# failed" messages usually also contain "[SSL:" which would otherwise
-# match the transient list and retry forever.
-_SSL_CERT_VERIFY_PATTERNS = [
-    "certificate verify failed",       # Python ssl module canonical text
-    "certificate_verify_failed",       # OpenSSL error token
-    "unable to get local issuer certificate",
-    "self-signed certificate",
-    "self signed certificate",
-    "certificate has expired",
-    "hostname mismatch, certificate is not valid",
-    "unable to verify the first certificate",  # Node/undici phrasing (MCP bridges)
-]
-
 # SSL/TLS transient failure patterns — intentionally distinct from
 # _SERVER_DISCONNECT_PATTERNS above.
 #
@@ -794,14 +680,6 @@ def classify_api_error(
         if classified is not None:
             return classified
 
-    # Local MoA config drift is deterministic: a persisted session can retain
-    # a preset name that was later renamed/deleted. Retrying the same lookup
-    # cannot recover and makes a clear config error look like an API outage.
-    from agent.errors import MoAPresetNotFoundError
-
-    if isinstance(error, MoAPresetNotFoundError):
-        return _result(FailoverReason.model_not_found, retryable=False)
-
     # ── 3. Error code classification ────────────────────────────────
 
     if error_code:
@@ -820,22 +698,7 @@ def classify_api_error(
     if classified is not None:
         return classified
 
-    # ── 5. SSL certificate verification failures → fail fast ────────
-    # A broken certificate chain (TLS-inspecting proxy, missing custom CA,
-    # expired/self-signed cert) is deterministic for the host — every retry
-    # reproduces the identical handshake failure. Fail immediately with
-    # actionable guidance instead of burning the retry budget first.
-    # Checked BEFORE the transient-SSL patterns: cert-verify messages also
-    # contain "[ssl:" which would otherwise match the transient list.
-    # Inspired by Claude Code v2.1.199 (July 2026).
-    if any(p in error_msg for p in _SSL_CERT_VERIFY_PATTERNS):
-        return _result(
-            FailoverReason.ssl_cert_verification,
-            retryable=False,
-            should_fallback=False,
-        )
-
-    # ── 5b. SSL/TLS transient errors → retry as timeout (not compression) ──
+    # ── 5. SSL/TLS transient errors → retry as timeout (not compression) ──
     # SSL alerts mid-stream are transport hiccups, not server-side context
     # overflow signals.  Classify before the disconnect check so a large
     # session doesn't incorrectly trigger context compression when the real
@@ -854,26 +717,6 @@ def classify_api_error(
 
     is_disconnect = any(p in error_msg for p in _SERVER_DISCONNECT_PATTERNS)
     if is_disconnect and not status_code:
-        # Reasoning-model override: a transport disconnect on a reasoning
-        # model is much more likely the upstream proxy idle-killing a
-        # long thinking stream than a true context overflow — even on
-        # large sessions.  The default disconnect+large-session routing
-        # below would otherwise send the user into the compression
-        # branch (should_compress=True) and silently delete
-        # conversation history on a phantom context-length error.
-        # Reasoning models have multi-minute thinking phases that
-        # routinely exceed the cloud gateway's idle window (NVIDIA
-        # NIM ~120s — first-party repro at NVIDIA/NemoClaw#4846;
-        # OpenAI worker / Anthropic stream-idle similar).  The
-        # per-reasoning-model stale-timeout floor in
-        # agent/reasoning_timeouts.py raises the stale-detector
-        # threshold to tolerate long thinking, so a true
-        # transport-layer failure here is recoverable via the retry
-        # path — not via context compression.  Reclassify as timeout.
-        # (Part 1 of Fixes #52310.)
-        from agent.reasoning_timeouts import get_reasoning_stale_timeout_floor
-        if get_reasoning_stale_timeout_floor(model) is not None:
-            return _result(FailoverReason.timeout, retryable=True)
         # Absolute token/message-count thresholds are only a proxy for smaller
         # context windows.  Large-context sessions can have hundreds of
         # messages while still being far below their actual token budget.
@@ -888,34 +731,12 @@ def classify_api_error(
             )
         return _result(FailoverReason.timeout, retryable=True)
 
-    # ── 7b. Stale-call circuit breaker → failover immediately ──────
-    # _check_stale_giveup() in agent/chat_completion_helpers.py raises a
-    # RuntimeError when the provider has been unresponsive for N
-    # consecutive stale attempts (default 5).  The error is NOT a transport
-    # timeout — the circuit breaker fires *before* any network call to avoid
-    # an indefinite stall.  Without this classification the RuntimeError
-    # falls through to FailoverReason.unknown (retryable=True), which burns
-    # all max_retries against the same dead provider (each retry hitting the
-    # circuit breaker instantly with zero network overhead) before fallback
-    # is attempted.  Classify as non-retryable + should_fallback so the
-    # retry loop activates the next fallback provider on the first hit.
-    if (
-        error_type == "RuntimeError"
-        and "consecutive stale attempts" in error_msg
-        and "aborting this call" in error_msg
-    ):
-        return _result(
-            FailoverReason.timeout,
-            retryable=False,
-            should_fallback=True,
-        )
-
-    # ── 8. Transport / timeout heuristics ───────────────────────────
+    # ── 7. Transport / timeout heuristics ───────────────────────────
 
     if error_type in _TRANSPORT_ERROR_TYPES or isinstance(error, (TimeoutError, ConnectionError, OSError)):
         return _result(FailoverReason.timeout, retryable=True)
 
-    # ── 9. Fallback: unknown ────────────────────────────────────────
+    # ── 8. Fallback: unknown ────────────────────────────────────────
 
     return _result(FailoverReason.unknown, retryable=True)
 
@@ -954,11 +775,7 @@ def _classify_by_status(
         # OpenRouter 403 "key limit exceeded" is actually billing. Other
         # providers also use 403 for account-plan or credit exhaustion.
         if (
-            (
-                provider == "xai-oauth"
-                and error_code.lower() == _XAI_SPENDING_LIMIT_ERROR_CODE
-            )
-            or "key limit exceeded" in error_msg
+            "key limit exceeded" in error_msg
             or "spending limit" in error_msg
             or any(p in error_msg for p in _BILLING_PATTERNS)
         ):
@@ -1026,35 +843,7 @@ def _classify_by_status(
         )
 
     if status_code == 429:
-        # Already checked long_context_tier above. Some providers (notably
-        # Z.AI / Zhipu) reuse HTTP 429 for server-wide overload — same status
-        # code as a true per-credential rate limit, but the credential is
-        # valid and the correct recovery is "back off and retry the same key",
-        # NOT "rotate the credential" (which exhausts the pool while the
-        # endpoint is still busy, and does nothing for a single-key user).
-        # Disambiguate on the error body so an overload 429 takes the
-        # transient-overload path instead of burning the pool. (#14038)
-        if any(p in error_msg for p in _OVERLOADED_PATTERNS):
-            return result_fn(
-                FailoverReason.overloaded,
-                retryable=True,
-            )
-        # Distinguish an OpenRouter-aggregator upstream 429 (an upstream model
-        # like DeepSeek rate-limited OpenRouter's aggregate traffic) from an
-        # account-level 429 (the user's key is actually throttled). OpenRouter
-        # wraps upstream errors with the outer message "Provider returned
-        # error" — the user's key is healthy, so marking it exhausted / rotating
-        # is wrong and burns the key for ~24min. Fall back to a different model.
-        if _is_openrouter_upstream_error(body, provider):
-            upstream_provider = _extract_upstream_provider_name(body)
-            ctx = {"upstream_provider": upstream_provider} if upstream_provider else {}
-            return result_fn(
-                FailoverReason.upstream_rate_limit,
-                retryable=True,
-                should_rotate_credential=False,
-                should_fallback=True,
-                error_context=ctx,
-            )
+        # Already checked long_context_tier above; this is a normal rate limit
         return result_fn(
             FailoverReason.rate_limit,
             retryable=True,
@@ -1090,58 +879,11 @@ def _classify_by_status(
                 retryable=False,
                 should_fallback=True,
             )
-        # Some local inference servers (notably llama.cpp / llama-server)
-        # report context overflow with an HTTP 500 instead of the standard
-        # 400/413. The request-validation guard above already ran, so any
-        # remaining explicit context-overflow signal routes into the
-        # compression-and-retry path (mirroring _classify_400) instead of
-        # blind server_error retries that exhaust and drop the turn.
-        # Empty-response advisories that mention "max_tokens" must not enter
-        # that compression path.
-        if any(p in error_msg for p in _EMPTY_PROVIDER_RESPONSE_PATTERNS):
-            return result_fn(
-                FailoverReason.server_error,
-                retryable=True,
-                should_compress=False,
-            )
-        if any(p in error_msg for p in _CONTEXT_OVERFLOW_PATTERNS):
-            return result_fn(
-                FailoverReason.context_overflow,
-                retryable=True,
-                should_compress=True,
-            )
         return result_fn(FailoverReason.server_error, retryable=True)
 
     if status_code in {503, 529}:
-        # Same overflow-as-5xx variant (server busy / model-load OOM, or a
-        # Cloudflare/Tailscale hop relabeling the status). Route explicit
-        # overflow bodies into compression; otherwise treat as transient
-        # overload and retry.
-        if any(p in error_msg for p in _EMPTY_PROVIDER_RESPONSE_PATTERNS):
-            return result_fn(
-                FailoverReason.server_error,
-                retryable=True,
-                should_compress=False,
-            )
-        if any(p in error_msg for p in _CONTEXT_OVERFLOW_PATTERNS):
-            return result_fn(
-                FailoverReason.context_overflow,
-                retryable=True,
-                should_compress=True,
-            )
         return result_fn(FailoverReason.overloaded, retryable=True)
 
-    # 408 Request Timeout — a transient timing failure the server itself flags
-    # as safe to retry (RFC 9110 §15.5.9), not a malformed request. Commonly
-    # emitted by reverse proxies sitting in front of self-hosted backends
-    # (llama.cpp / Ollama / vLLM) when a long generation outruns the proxy's
-    # request-read window. Route to the dedicated ``timeout`` reason (rebuild
-    # client + retry) instead of falling through to the generic 4xx bucket
-    # below, which would abort the turn on a retry-safe error the same way it
-    # aborts a 400 Bad Request.
-    if status_code == 408:
-        return result_fn(FailoverReason.timeout, retryable=True)
-
     # Other 4xx — non-retryable
     if 400 <= status_code < 500:
         return result_fn(
@@ -1235,7 +977,6 @@ def _classify_400(
             "encrypted content for item" in error_msg
             and "could not be verified" in error_msg
         )
-        or "could not decrypt the provided encrypted_content" in error_msg
     ):
         return result_fn(
             FailoverReason.invalid_encrypted_content,
@@ -1248,8 +989,8 @@ def _classify_400(
     # returns:
     #   "Unsupported parameter: 'max_tokens' is not supported with this model.
     #    Use 'max_completion_tokens' instead."
-    # That string contains the literal substring "max_tokens", which historically
-    # sat in _CONTEXT_OVERFLOW_PATTERNS — so without this guard the 400 is
+    # That string contains the literal substring "max_tokens", which is one of
+    # the _CONTEXT_OVERFLOW_PATTERNS — so without this guard the 400 is
     # misclassified as context_overflow, routed into the compression loop,
     # re-sent with the same bad parameter, and ends in "Cannot compress
     # further".  These errors are deterministic (every retry gets the identical
@@ -1271,17 +1012,6 @@ def _classify_400(
             should_fallback=True,
         )
 
-    # Empty-provider-response advisories must not enter compression. They
-    # often mention "max_tokens" as a possible cause and used to match the
-    # bare overflow pattern, then thrash compress until "Cannot compress
-    # further" on an otherwise healthy session (custom endpoints / nano-gpt).
-    if any(p in error_msg for p in _EMPTY_PROVIDER_RESPONSE_PATTERNS):
-        return result_fn(
-            FailoverReason.server_error,
-            retryable=True,
-            should_compress=False,
-        )
-
     # Context overflow from 400
     if any(p in error_msg for p in _CONTEXT_OVERFLOW_PATTERNS):
         return result_fn(
@@ -1369,7 +1099,15 @@ def _classify_by_error_code(
             should_rotate_credential=True,
         )
 
-    if code_lower in _BILLING_ERROR_CODES:
+    if code_lower in {
+        "insufficient_quota",
+        "billing_not_active",
+        "payment_required",
+        "insufficient_credits",
+        "no_usable_credits",
+        "balance_depleted",
+        "model_not_supported_on_free_tier",
+    }:
         return result_fn(
             FailoverReason.billing,
             retryable=False,
@@ -1456,17 +1194,6 @@ def _classify_by_message(
             should_fallback=True,
         )
 
-    # Overloaded / server-busy patterns — must come BEFORE the rate_limit and
-    # billing checks so that a message-only "overloaded" (no 503/529 status,
-    # e.g. some Anthropic-compatible proxies) classifies as a transient
-    # overload (backoff + retry) instead of falling through to `unknown` or
-    # incorrectly triggering credential rotation.
-    if any(p in error_msg for p in _OVERLOADED_PATTERNS):
-        return result_fn(
-            FailoverReason.overloaded,
-            retryable=True,
-        )
-
     # Billing patterns
     if any(p in error_msg for p in _BILLING_PATTERNS):
         return result_fn(
@@ -1485,15 +1212,6 @@ def _classify_by_message(
             should_fallback=True,
         )
 
-    # Empty-provider-response advisories (often mention "max_tokens") must
-    # retry without compression — see the matching 400-path guard above.
-    if any(p in error_msg for p in _EMPTY_PROVIDER_RESPONSE_PATTERNS):
-        return result_fn(
-            FailoverReason.server_error,
-            retryable=True,
-            should_compress=False,
-        )
-
     # Context overflow patterns
     if any(p in error_msg for p in _CONTEXT_OVERFLOW_PATTERNS):
         return result_fn(
@@ -1565,25 +1283,19 @@ def _extract_status_code(error: Exception) -> Optional[int]:
 
 
 def _extract_error_body(error: Exception) -> dict:
-    """Extract the structured error body from an SDK exception or its cause chain."""
-    current = error
-    for _ in range(5):  # Match _extract_status_code() traversal depth.
-        body = getattr(current, "body", None)
-        if isinstance(body, dict):
-            return body
-        # Some errors have .response.json()
-        response = getattr(current, "response", None)
-        if response is not None:
-            try:
-                json_body = response.json()
-                if isinstance(json_body, dict):
-                    return json_body
-            except Exception:
-                pass
-        cause = getattr(current, "__cause__", None) or getattr(current, "__context__", None)
-        if cause is None or cause is current:
-            break
-        current = cause
+    """Extract the structured error body from an SDK exception."""
+    body = getattr(error, "body", None)
+    if isinstance(body, dict):
+        return body
+    # Some errors have .response.json()
+    response = getattr(error, "response", None)
+    if response is not None:
+        try:
+            json_body = response.json()
+            if isinstance(json_body, dict):
+                return json_body
+        except Exception:
+            pass
     return {}
 
 
@@ -1651,49 +1363,3 @@ def _extract_message(error: Exception, body: dict) -> str:
             return msg.strip()[:500]
     # Fallback to str(error)
     return str(error)[:500]
-
-
-def _is_openrouter_upstream_error(body: Any, provider: str) -> bool:
-    """Detect OpenRouter's aggregator-wrapped upstream provider errors.
-
-    OpenRouter returns errors from upstream model providers (DeepSeek,
-    Anthropic, etc.) wrapped with the outer message "Provider returned error"
-    and the real error nested in ``metadata.raw``. This signal means the
-    user's OpenRouter key is healthy — the upstream provider is the one that
-    failed — so credential rotation is the wrong recovery.
-    """
-    if not isinstance(body, dict):
-        return False
-    provider_lower = (provider or "").strip().lower()
-    err = body.get("error")
-    if not isinstance(err, dict):
-        return False
-    outer_msg = str(err.get("message") or "").strip().lower()
-    if outer_msg != "provider returned error":
-        return False
-    # Require either the explicit OpenRouter provider OR the metadata shape
-    # that only OpenRouter produces (metadata.raw / metadata.provider_name).
-    if provider_lower == "openrouter":
-        return True
-    metadata = err.get("metadata")
-    if isinstance(metadata, dict) and (
-        "raw" in metadata or "provider_name" in metadata
-    ):
-        return True
-    return False
-
-
-def _extract_upstream_provider_name(body: Any) -> Optional[str]:
-    """Pull the upstream provider name out of OpenRouter's error metadata."""
-    if not isinstance(body, dict):
-        return None
-    err = body.get("error")
-    if not isinstance(err, dict):
-        return None
-    metadata = err.get("metadata")
-    if not isinstance(metadata, dict):
-        return None
-    name = metadata.get("provider_name")
-    if isinstance(name, str) and name.strip():
-        return name.strip()
-    return None

agent/errors.py

@@ -1,13 +1,3 @@
 class SSLConfigurationError(Exception):
     """Raised when SSL/TLS certificate bundle configuration fails."""
     pass
-
-
-class EmptyStreamError(RuntimeError):
-    """Raised when a provider closes a stream without yielding a response."""
-
-    pass
-
-
-class MoAPresetNotFoundError(ValueError):
-    """Raised when a persisted MoA preset no longer exists in config."""

agent/file_safety.py

@@ -46,9 +46,6 @@ def build_write_denied_paths(home: str) -> set[str]:
             # Top-level Anthropic PKCE credential store remains sensitive even
             # when a profile is active; default/non-profile sessions still read it.
             str(hermes_root / ".anthropic_oauth.json"),
-            # Bitwarden Secrets Manager encrypted disk cache.
-            str(hermes_home / "cache" / "bws_cache.enc.json"),
-            str(hermes_root / "cache" / "bws_cache.enc.json"),
             os.path.join(home, ".netrc"),
             os.path.join(home, ".pgpass"),
             os.path.join(home, ".npmrc"),
@@ -80,34 +77,27 @@ def build_write_denied_prefixes(home: str) -> list[str]:
     ]
 
 
-def get_safe_write_roots() -> set[str]:
-    """Return resolved HERMES_WRITE_SAFE_ROOT paths. Supports multiple directories
-    separated by ``os.pathsep`` (``:`` on Unix, ``;`` on Windows).
-    E.g., ``/opt/data:/var/www/html`` on Unix, ``C:\\data;D:\\www`` on Windows."""
-    env = os.getenv("HERMES_WRITE_SAFE_ROOT", "")
-    if not env:
-        return set()
-    roots: set[str] = set()
-    for path in env.split(os.pathsep):
-        if path:
-            try:
-                resolved = os.path.realpath(os.path.expanduser(path))
-                roots.add(resolved)
-            except (OSError, ValueError):
-                continue
-    return roots
+def get_safe_write_root() -> Optional[str]:
+    """Return the resolved HERMES_WRITE_SAFE_ROOT path, or None if unset."""
+    root = os.getenv("HERMES_WRITE_SAFE_ROOT", "")
+    if not root:
+        return None
+    try:
+        return os.path.realpath(os.path.expanduser(root))
+    except Exception:
+        return None
 
 
-def _classify_write_denial(path: str) -> Optional[str]:
-    """Return ``'credential'``, ``'safe_root'``, or ``None`` if writes are allowed."""
+def is_write_denied(path: str) -> bool:
+    """Return True if path is blocked by the write denylist or safe root."""
     home = os.path.realpath(os.path.expanduser("~"))
     resolved = os.path.realpath(os.path.expanduser(str(path)))
 
     if resolved in build_write_denied_paths(home):
-        return "credential"
+        return True
     for prefix in build_write_denied_prefixes(home):
         if resolved.startswith(prefix):
-            return "credential"
+            return True
 
     mcp_tokens_dir_name = "mcp-tokens"
 
@@ -121,60 +111,24 @@ def _classify_write_denial(path: str) -> Optional[str]:
             continue
 
     for base_real in hermes_dirs:
-        # Session transcripts are application-owned state.  Letting the agent's
-        # generic file tools rewrite state.db or legacy JSON snapshots can
-        # falsify conversation history and invalidate resume/compression state.
-        try:
-            if resolved == os.path.realpath(os.path.join(base_real, "state.db")):
-                return True
-            sessions_real = os.path.realpath(os.path.join(base_real, "sessions"))
-            if resolved == sessions_real or resolved.startswith(sessions_real + os.sep):
-                return True
-        except Exception:
-            pass
         try:
             mcp_real = os.path.realpath(os.path.join(base_real, mcp_tokens_dir_name))
             if resolved == mcp_real or resolved.startswith(mcp_real + os.sep):
-                return "credential"
+                return True
         except Exception:
             pass
         try:
             pairing_real = os.path.realpath(os.path.join(base_real, "pairing"))
             if resolved == pairing_real or resolved.startswith(pairing_real + os.sep):
-                return "credential"
+                return True
         except Exception:
             pass
 
-    safe_roots = get_safe_write_roots()
-    if safe_roots:
-        allowed = False
-        for safe_root in safe_roots:
-            if resolved == safe_root or resolved.startswith(safe_root + os.sep):
-                allowed = True
-                break
-        if not allowed:
-            return "safe_root"
+    safe_root = get_safe_write_root()
+    if safe_root and not (resolved == safe_root or resolved.startswith(safe_root + os.sep)):
+        return True
 
-    return None
-
-
-def is_write_denied(path: str) -> bool:
-    """Return True if path is blocked by the write denylist or safe root."""
-    return _classify_write_denial(path) is not None
-
-
-def get_write_denied_error(path: str, *, verb: str = "Write") -> Optional[str]:
-    """Return a user/model-facing error when writes to ``path`` are blocked."""
-    denial = _classify_write_denial(path)
-    if denial is None:
-        return None
-    if denial == "safe_root":
-        roots_display = os.pathsep.join(sorted(get_safe_write_roots()))
-        return (
-            f"{verb} denied: '{path}' is outside HERMES_WRITE_SAFE_ROOT "
-            f"({roots_display}). Unset the variable or add this path's directory prefix."
-        )
-    return f"{verb} denied: '{path}' is a protected system/credential file."
+    return False
 
 
 # Common secret-bearing project-local environment file basenames.
@@ -326,7 +280,7 @@ def get_read_block_error(path: str) -> Optional[str]:
     # .env contents — .env.example is the documented-shape substitute. The
     # terminal tool can still ``cat .env``; this is defense-in-depth, not a
     # boundary (see module docstring).
-    if resolved.name.lower() in _BLOCKED_PROJECT_ENV_BASENAMES:
+    if resolved.name in _BLOCKED_PROJECT_ENV_BASENAMES:
         return (
             f"Access denied: {path} is a secret-bearing environment file "
             "and cannot be read to prevent credential leakage. "
@@ -337,30 +291,6 @@ def get_read_block_error(path: str) -> Optional[str]:
     return None
 
 
-def raise_if_read_blocked(path: str) -> None:
-    """Raise ``ValueError`` if ``path`` is a denied Hermes read (see
-    :func:`get_read_block_error`), else return.
-
-    Shared chokepoint for provider input-loading sites that read a local
-    file the model/tool supplied (e.g. image-gen ``image_url`` /
-    ``reference_image_urls`` paths). Centralizes the guard so every provider
-    enforces the same read boundary with identical semantics instead of each
-    open-coding the try/except block (#57698).
-
-    Best-effort by design: if ``agent.file_safety`` machinery is somehow
-    unavailable at the call site the guard no-ops rather than breaking local
-    image loading — consistent with the defense-in-depth (not security
-    boundary) framing of the denylist itself. The blocking ``ValueError`` from
-    a real hit still propagates; only unexpected internal errors are swallowed.
-    """
-    try:
-        blocked = get_read_block_error(path)
-    except Exception:  # noqa: BLE001 - guard must never break local-file loading
-        return
-    if blocked:
-        raise ValueError(blocked)
-
-
 # ---------------------------------------------------------------------------
 # Cross-profile write guard (#TBD)
 #

agent/gemini_cloudcode_adapter.py

@@ -0,0 +1,915 @@
+"""OpenAI-compatible facade that talks to Google's Cloud Code Assist backend.
+
+This adapter lets Hermes use the ``google-gemini-cli`` provider as if it were
+a standard OpenAI-shaped chat completion endpoint, while the underlying HTTP
+traffic goes to ``cloudcode-pa.googleapis.com/v1internal:{generateContent,
+streamGenerateContent}`` with a Bearer access token obtained via OAuth PKCE.
+
+Architecture
+------------
+- ``GeminiCloudCodeClient`` exposes ``.chat.completions.create(**kwargs)``
+  mirroring the subset of the OpenAI SDK that ``run_agent.py`` uses.
+- Incoming OpenAI ``messages[]`` / ``tools[]`` / ``tool_choice`` are translated
+  to Gemini's native ``contents[]`` / ``tools[].functionDeclarations`` /
+  ``toolConfig`` / ``systemInstruction`` shape.
+- The request body is wrapped ``{project, model, user_prompt_id, request}``
+  per Code Assist API expectations.
+- Responses (``candidates[].content.parts[]``) are converted back to
+  OpenAI ``choices[0].message`` shape with ``content`` + ``tool_calls``.
+- Streaming uses SSE (``?alt=sse``) and yields OpenAI-shaped delta chunks.
+
+Attribution
+-----------
+Translation semantics follow jenslys/opencode-gemini-auth (MIT) and the public
+Gemini API docs. Request envelope shape
+(``{project, model, user_prompt_id, request}``) is documented nowhere; it is
+reverse-engineered from the opencode-gemini-auth and clawdbot implementations.
+"""
+
+from __future__ import annotations
+
+import json
+import logging
+import time
+import uuid
+from types import SimpleNamespace
+from typing import Any, Dict, Iterator, List, Optional
+
+import httpx
+
+from agent import google_oauth
+from agent.gemini_schema import sanitize_gemini_tool_parameters
+from agent.google_code_assist import (
+    CODE_ASSIST_ENDPOINT,
+    CodeAssistError,
+    ProjectContext,
+    resolve_project_context,
+)
+
+logger = logging.getLogger(__name__)
+
+
+# =============================================================================
+# Request translation: OpenAI → Gemini
+# =============================================================================
+
+_ROLE_MAP_OPENAI_TO_GEMINI = {
+    "user": "user",
+    "assistant": "model",
+    "system": "user",   # handled separately via systemInstruction
+    "tool": "user",     # functionResponse is wrapped in a user-role turn
+    "function": "user",
+}
+
+
+def _coerce_content_to_text(content: Any) -> str:
+    """OpenAI content may be str or a list of parts; reduce to plain text."""
+    if content is None:
+        return ""
+    if isinstance(content, str):
+        return content
+    if isinstance(content, list):
+        pieces: List[str] = []
+        for p in content:
+            if isinstance(p, str):
+                pieces.append(p)
+            elif isinstance(p, dict):
+                if p.get("type") == "text" and isinstance(p.get("text"), str):
+                    pieces.append(p["text"])
+                # Multimodal (image_url, etc.) — stub for now; log and skip
+                elif p.get("type") in {"image_url", "input_audio"}:
+                    logger.debug("Dropping multimodal part (not yet supported): %s", p.get("type"))
+        return "\n".join(pieces)
+    return str(content)
+
+
+def _translate_tool_call_to_gemini(tool_call: Dict[str, Any]) -> Dict[str, Any]:
+    """OpenAI tool_call -> Gemini functionCall part."""
+    fn = tool_call.get("function") or {}
+    args_raw = fn.get("arguments", "")
+    try:
+        args = json.loads(args_raw) if isinstance(args_raw, str) and args_raw else {}
+    except json.JSONDecodeError:
+        args = {"_raw": args_raw}
+    if not isinstance(args, dict):
+        args = {"_value": args}
+    function_call = {
+        "name": fn.get("name") or "",
+        "args": args,
+    }
+    if tool_call.get("id"):
+        function_call["id"] = str(tool_call["id"])
+    return {
+        "functionCall": function_call,
+        # Sentinel signature — matches opencode-gemini-auth's approach.
+        # Without this, Code Assist rejects function calls that originated
+        # outside its own chain.
+        "thoughtSignature": "skip_thought_signature_validator",
+    }
+
+
+def _translate_tool_result_to_gemini(message: Dict[str, Any]) -> Dict[str, Any]:
+    """OpenAI tool-role message -> Gemini functionResponse part.
+
+    The function name isn't in the OpenAI tool message directly; it must be
+    passed via the assistant message that issued the call. For simplicity we
+    look up ``name`` on the message (OpenAI SDK copies it there) or on the
+    ``tool_call_id`` cross-reference.
+    """
+    name = str(message.get("name") or message.get("tool_call_id") or "tool")
+    content = _coerce_content_to_text(message.get("content"))
+    # Gemini expects the response as a dict under `response`. We wrap plain
+    # text in {"output": "..."}.
+    try:
+        parsed = json.loads(content) if content.strip().startswith(("{", "[")) else None
+    except json.JSONDecodeError:
+        parsed = None
+    response = parsed if isinstance(parsed, dict) else {"output": content}
+    function_response = {
+        "name": name,
+        "response": response,
+    }
+    if message.get("tool_call_id"):
+        function_response["id"] = str(message["tool_call_id"])
+    return {"functionResponse": function_response}
+
+
+def _build_gemini_contents(
+    messages: List[Dict[str, Any]],
+) -> tuple[List[Dict[str, Any]], Optional[Dict[str, Any]]]:
+    """Convert OpenAI messages[] to Gemini contents[] + systemInstruction."""
+    system_text_parts: List[str] = []
+    contents: List[Dict[str, Any]] = []
+
+    for msg in messages:
+        if not isinstance(msg, dict):
+            continue
+        role = str(msg.get("role") or "user")
+
+        if role == "system":
+            system_text_parts.append(_coerce_content_to_text(msg.get("content")))
+            continue
+
+        # Tool result message — emit a user-role turn with functionResponse
+        if role == "tool" or role == "function":
+            contents.append({
+                "role": "user",
+                "parts": [_translate_tool_result_to_gemini(msg)],
+            })
+            continue
+
+        gemini_role = _ROLE_MAP_OPENAI_TO_GEMINI.get(role, "user")
+        parts: List[Dict[str, Any]] = []
+
+        text = _coerce_content_to_text(msg.get("content"))
+        if text:
+            parts.append({"text": text})
+
+        # Assistant messages can carry tool_calls
+        tool_calls = msg.get("tool_calls") or []
+        if isinstance(tool_calls, list):
+            for tc in tool_calls:
+                if isinstance(tc, dict):
+                    parts.append(_translate_tool_call_to_gemini(tc))
+
+        if not parts:
+            # Gemini rejects empty parts; skip the turn entirely
+            continue
+
+        contents.append({"role": gemini_role, "parts": parts})
+
+    system_instruction: Optional[Dict[str, Any]] = None
+    joined_system = "\n".join(p for p in system_text_parts if p).strip()
+    if joined_system:
+        system_instruction = {
+            "role": "system",
+            "parts": [{"text": joined_system}],
+        }
+
+    return contents, system_instruction
+
+
+def _translate_tools_to_gemini(tools: Any) -> List[Dict[str, Any]]:
+    """OpenAI tools[] -> Gemini tools[].functionDeclarations[]."""
+    if not isinstance(tools, list) or not tools:
+        return []
+    declarations: List[Dict[str, Any]] = []
+    for t in tools:
+        if not isinstance(t, dict):
+            continue
+        fn = t.get("function") or {}
+        if not isinstance(fn, dict):
+            continue
+        name = fn.get("name")
+        if not name:
+            continue
+        decl = {"name": str(name)}
+        if fn.get("description"):
+            decl["description"] = str(fn["description"])
+        params = fn.get("parameters")
+        if isinstance(params, dict):
+            decl["parameters"] = sanitize_gemini_tool_parameters(params)
+        declarations.append(decl)
+    if not declarations:
+        return []
+    return [{"functionDeclarations": declarations}]
+
+
+def _translate_tool_choice_to_gemini(tool_choice: Any) -> Optional[Dict[str, Any]]:
+    """OpenAI tool_choice -> Gemini toolConfig.functionCallingConfig."""
+    if tool_choice is None:
+        return None
+    if isinstance(tool_choice, str):
+        if tool_choice == "auto":
+            return {"functionCallingConfig": {"mode": "AUTO"}}
+        if tool_choice == "required":
+            return {"functionCallingConfig": {"mode": "ANY"}}
+        if tool_choice == "none":
+            return {"functionCallingConfig": {"mode": "NONE"}}
+    if isinstance(tool_choice, dict):
+        fn = tool_choice.get("function") or {}
+        name = fn.get("name")
+        if name:
+            return {
+                "functionCallingConfig": {
+                    "mode": "ANY",
+                    "allowedFunctionNames": [str(name)],
+                },
+            }
+    return None
+
+
+def _normalize_thinking_config(config: Any) -> Optional[Dict[str, Any]]:
+    """Accept thinkingBudget / thinkingLevel / includeThoughts (+ snake_case)."""
+    if not isinstance(config, dict) or not config:
+        return None
+    budget = config.get("thinkingBudget", config.get("thinking_budget"))
+    level = config.get("thinkingLevel", config.get("thinking_level"))
+    include = config.get("includeThoughts", config.get("include_thoughts"))
+    normalized: Dict[str, Any] = {}
+    if isinstance(budget, (int, float)):
+        normalized["thinkingBudget"] = int(budget)
+    if isinstance(level, str) and level.strip():
+        normalized["thinkingLevel"] = level.strip().lower()
+    if isinstance(include, bool):
+        normalized["includeThoughts"] = include
+    return normalized or None
+
+
+def build_gemini_request(
+    *,
+    messages: List[Dict[str, Any]],
+    tools: Any = None,
+    tool_choice: Any = None,
+    temperature: Optional[float] = None,
+    max_tokens: Optional[int] = None,
+    top_p: Optional[float] = None,
+    stop: Any = None,
+    thinking_config: Any = None,
+) -> Dict[str, Any]:
+    """Build the inner Gemini request body (goes inside ``request`` wrapper)."""
+    contents, system_instruction = _build_gemini_contents(messages)
+
+    body: Dict[str, Any] = {"contents": contents}
+    if system_instruction is not None:
+        body["systemInstruction"] = system_instruction
+
+    gemini_tools = _translate_tools_to_gemini(tools)
+    if gemini_tools:
+        body["tools"] = gemini_tools
+    tool_cfg = _translate_tool_choice_to_gemini(tool_choice)
+    if tool_cfg is not None:
+        body["toolConfig"] = tool_cfg
+
+    generation_config: Dict[str, Any] = {}
+    if isinstance(temperature, (int, float)):
+        generation_config["temperature"] = float(temperature)
+    if isinstance(max_tokens, int) and max_tokens > 0:
+        generation_config["maxOutputTokens"] = max_tokens
+    if isinstance(top_p, (int, float)):
+        generation_config["topP"] = float(top_p)
+    if isinstance(stop, str) and stop:
+        generation_config["stopSequences"] = [stop]
+    elif isinstance(stop, list) and stop:
+        generation_config["stopSequences"] = [str(s) for s in stop if s]
+    normalized_thinking = _normalize_thinking_config(thinking_config)
+    if normalized_thinking:
+        generation_config["thinkingConfig"] = normalized_thinking
+    if generation_config:
+        body["generationConfig"] = generation_config
+
+    return body
+
+
+def wrap_code_assist_request(
+    *,
+    project_id: str,
+    model: str,
+    inner_request: Dict[str, Any],
+    user_prompt_id: Optional[str] = None,
+) -> Dict[str, Any]:
+    """Wrap the inner Gemini request in the Code Assist envelope."""
+    return {
+        "project": project_id,
+        "model": model,
+        "user_prompt_id": user_prompt_id or str(uuid.uuid4()),
+        "request": inner_request,
+    }
+
+
+# =============================================================================
+# Response translation: Gemini → OpenAI
+# =============================================================================
+
+def _translate_gemini_response(
+    resp: Dict[str, Any],
+    model: str,
+) -> SimpleNamespace:
+    """Non-streaming Gemini response -> OpenAI-shaped SimpleNamespace.
+
+    Code Assist wraps the actual Gemini response inside ``response``, so we
+    unwrap it first if present.
+    """
+    inner = resp.get("response") if isinstance(resp.get("response"), dict) else resp
+
+    candidates = inner.get("candidates") or []
+    if not isinstance(candidates, list) or not candidates:
+        return _empty_response(model)
+
+    cand = candidates[0]
+    content_obj = cand.get("content") if isinstance(cand, dict) else {}
+    parts = content_obj.get("parts") if isinstance(content_obj, dict) else []
+
+    text_pieces: List[str] = []
+    reasoning_pieces: List[str] = []
+    tool_calls: List[SimpleNamespace] = []
+
+    for i, part in enumerate(parts or []):
+        if not isinstance(part, dict):
+            continue
+        # Thought parts are model's internal reasoning — surface as reasoning,
+        # don't mix into content.
+        if part.get("thought") is True:
+            if isinstance(part.get("text"), str):
+                reasoning_pieces.append(part["text"])
+            continue
+        if isinstance(part.get("text"), str):
+            text_pieces.append(part["text"])
+            continue
+        fc = part.get("functionCall")
+        if isinstance(fc, dict) and fc.get("name"):
+            try:
+                args_str = json.dumps(fc.get("args") or {}, ensure_ascii=False)
+            except (TypeError, ValueError):
+                args_str = "{}"
+            call_id = str(fc.get("id") or "").strip() or f"call_{uuid.uuid4().hex[:12]}"
+            tool_calls.append(SimpleNamespace(
+                id=call_id,
+                type="function",
+                index=i,
+                function=SimpleNamespace(name=str(fc["name"]), arguments=args_str),
+            ))
+
+    finish_reason = "tool_calls" if tool_calls else _map_gemini_finish_reason(
+        str(cand.get("finishReason") or "")
+    )
+
+    usage_meta = inner.get("usageMetadata") or {}
+    usage = SimpleNamespace(
+        prompt_tokens=int(usage_meta.get("promptTokenCount") or 0),
+        completion_tokens=int(usage_meta.get("candidatesTokenCount") or 0),
+        total_tokens=int(usage_meta.get("totalTokenCount") or 0),
+        prompt_tokens_details=SimpleNamespace(
+            cached_tokens=int(usage_meta.get("cachedContentTokenCount") or 0),
+        ),
+    )
+
+    message = SimpleNamespace(
+        role="assistant",
+        content="".join(text_pieces) if text_pieces else None,
+        tool_calls=tool_calls or None,
+        reasoning="".join(reasoning_pieces) or None,
+        reasoning_content="".join(reasoning_pieces) or None,
+        reasoning_details=None,
+    )
+    choice = SimpleNamespace(
+        index=0,
+        message=message,
+        finish_reason=finish_reason,
+    )
+    return SimpleNamespace(
+        id=f"chatcmpl-{uuid.uuid4().hex[:12]}",
+        object="chat.completion",
+        created=int(time.time()),
+        model=model,
+        choices=[choice],
+        usage=usage,
+    )
+
+
+def _empty_response(model: str) -> SimpleNamespace:
+    message = SimpleNamespace(
+        role="assistant", content="", tool_calls=None,
+        reasoning=None, reasoning_content=None, reasoning_details=None,
+    )
+    choice = SimpleNamespace(index=0, message=message, finish_reason="stop")
+    usage = SimpleNamespace(
+        prompt_tokens=0, completion_tokens=0, total_tokens=0,
+        prompt_tokens_details=SimpleNamespace(cached_tokens=0),
+    )
+    return SimpleNamespace(
+        id=f"chatcmpl-{uuid.uuid4().hex[:12]}",
+        object="chat.completion",
+        created=int(time.time()),
+        model=model,
+        choices=[choice],
+        usage=usage,
+    )
+
+
+def _map_gemini_finish_reason(reason: str) -> str:
+    mapping = {
+        "STOP": "stop",
+        "MAX_TOKENS": "length",
+        "SAFETY": "content_filter",
+        "RECITATION": "content_filter",
+        "OTHER": "stop",
+    }
+    return mapping.get(reason.upper(), "stop")
+
+
+# =============================================================================
+# Streaming SSE iterator
+# =============================================================================
+
+class _GeminiStreamChunk(SimpleNamespace):
+    """Mimics an OpenAI ChatCompletionChunk with .choices[0].delta."""
+    pass
+
+
+def _make_stream_chunk(
+    *,
+    model: str,
+    content: str = "",
+    tool_call_delta: Optional[Dict[str, Any]] = None,
+    finish_reason: Optional[str] = None,
+    reasoning: str = "",
+) -> _GeminiStreamChunk:
+    delta_kwargs: Dict[str, Any] = {
+        "role": "assistant",
+        "content": None,
+        "tool_calls": None,
+        "reasoning": None,
+        "reasoning_content": None,
+    }
+    if content:
+        delta_kwargs["content"] = content
+    if tool_call_delta is not None:
+        delta_kwargs["tool_calls"] = [SimpleNamespace(
+            index=tool_call_delta.get("index", 0),
+            id=tool_call_delta.get("id") or f"call_{uuid.uuid4().hex[:12]}",
+            type="function",
+            function=SimpleNamespace(
+                name=tool_call_delta.get("name") or "",
+                arguments=tool_call_delta.get("arguments") or "",
+            ),
+        )]
+    if reasoning:
+        delta_kwargs["reasoning"] = reasoning
+        delta_kwargs["reasoning_content"] = reasoning
+    delta = SimpleNamespace(**delta_kwargs)
+    choice = SimpleNamespace(index=0, delta=delta, finish_reason=finish_reason)
+    return _GeminiStreamChunk(
+        id=f"chatcmpl-{uuid.uuid4().hex[:12]}",
+        object="chat.completion.chunk",
+        created=int(time.time()),
+        model=model,
+        choices=[choice],
+        usage=None,
+    )
+
+
+def _iter_sse_events(response: httpx.Response) -> Iterator[Dict[str, Any]]:
+    """Parse Server-Sent Events from an httpx streaming response."""
+    buffer = ""
+    for chunk in response.iter_text():
+        if not chunk:
+            continue
+        buffer += chunk
+        while "\n" in buffer:
+            line, buffer = buffer.split("\n", 1)
+            line = line.rstrip("\r")
+            if not line:
+                continue
+            if line.startswith("data: "):
+                data = line[6:]
+                if data == "[DONE]":
+                    return
+                try:
+                    yield json.loads(data)
+                except json.JSONDecodeError:
+                    logger.debug("Non-JSON SSE line: %s", data[:200])
+
+
+def _translate_stream_event(
+    event: Dict[str, Any],
+    model: str,
+    tool_call_counter: List[int],
+) -> List[_GeminiStreamChunk]:
+    """Unwrap Code Assist envelope and emit OpenAI-shaped chunk(s).
+
+    ``tool_call_counter`` is a single-element list used as a mutable counter
+    across events in the same stream. Each ``functionCall`` part gets a
+    fresh, unique OpenAI ``index`` — keying by function name would collide
+    whenever the model issues parallel calls to the same tool (e.g. reading
+    three files in one turn).
+    """
+    inner = event.get("response") if isinstance(event.get("response"), dict) else event
+    candidates = inner.get("candidates") or []
+    if not candidates:
+        return []
+    cand = candidates[0]
+    if not isinstance(cand, dict):
+        return []
+
+    chunks: List[_GeminiStreamChunk] = []
+
+    content = cand.get("content") or {}
+    parts = content.get("parts") if isinstance(content, dict) else []
+    for part in parts or []:
+        if not isinstance(part, dict):
+            continue
+        if part.get("thought") is True and isinstance(part.get("text"), str):
+            chunks.append(_make_stream_chunk(
+                model=model, reasoning=part["text"],
+            ))
+            continue
+        if isinstance(part.get("text"), str) and part["text"]:
+            chunks.append(_make_stream_chunk(model=model, content=part["text"]))
+        fc = part.get("functionCall")
+        if isinstance(fc, dict) and fc.get("name"):
+            name = str(fc["name"])
+            idx = tool_call_counter[0]
+            tool_call_counter[0] += 1
+            try:
+                args_str = json.dumps(fc.get("args") or {}, ensure_ascii=False)
+            except (TypeError, ValueError):
+                args_str = "{}"
+            chunks.append(_make_stream_chunk(
+                model=model,
+                tool_call_delta={
+                    "index": idx,
+                    "id": str(fc.get("id") or "").strip(),
+                    "name": name,
+                    "arguments": args_str,
+                },
+            ))
+
+    finish_reason_raw = str(cand.get("finishReason") or "")
+    if finish_reason_raw:
+        mapped = _map_gemini_finish_reason(finish_reason_raw)
+        if tool_call_counter[0] > 0:
+            mapped = "tool_calls"
+        chunks.append(_make_stream_chunk(model=model, finish_reason=mapped))
+    return chunks
+
+
+# =============================================================================
+# GeminiCloudCodeClient — OpenAI-compatible facade
+# =============================================================================
+
+MARKER_BASE_URL = "cloudcode-pa://google"
+
+
+class _GeminiChatCompletions:
+    def __init__(self, client: "GeminiCloudCodeClient"):
+        self._client = client
+
+    def create(self, **kwargs: Any) -> Any:
+        return self._client._create_chat_completion(**kwargs)
+
+
+class _GeminiChatNamespace:
+    def __init__(self, client: "GeminiCloudCodeClient"):
+        self.completions = _GeminiChatCompletions(client)
+
+
+class GeminiCloudCodeClient:
+    """Minimal OpenAI-SDK-compatible facade over Code Assist v1internal."""
+
+    def __init__(
+        self,
+        *,
+        api_key: Optional[str] = None,
+        base_url: Optional[str] = None,
+        default_headers: Optional[Dict[str, str]] = None,
+        project_id: str = "",
+        **_: Any,
+    ):
+        # `api_key` here is a dummy — real auth is the OAuth access token
+        # fetched on every call via agent.google_oauth.get_valid_access_token().
+        # We accept the kwarg for openai.OpenAI interface parity.
+        self.api_key = api_key or "google-oauth"
+        self.base_url = base_url or MARKER_BASE_URL
+        self._default_headers = dict(default_headers or {})
+        self._configured_project_id = project_id
+        self._project_context: Optional[ProjectContext] = None
+        self._project_context_lock = False  # simple single-thread guard
+        self.chat = _GeminiChatNamespace(self)
+        self.is_closed = False
+        self._http = httpx.Client(timeout=httpx.Timeout(connect=15.0, read=600.0, write=30.0, pool=30.0))
+
+    def close(self) -> None:
+        self.is_closed = True
+        try:
+            self._http.close()
+        except Exception:
+            pass
+
+    # Implement the OpenAI SDK's context-manager-ish closure check
+    def __enter__(self):
+        return self
+
+    def __exit__(self, exc_type, exc_val, exc_tb):
+        self.close()
+
+    def _ensure_project_context(self, access_token: str, model: str) -> ProjectContext:
+        """Lazily resolve and cache the project context for this client."""
+        if self._project_context is not None:
+            return self._project_context
+
+        env_project = google_oauth.resolve_project_id_from_env()
+        creds = google_oauth.load_credentials()
+        stored_project = creds.project_id if creds else ""
+
+        # Prefer what's already baked into the creds
+        if stored_project:
+            self._project_context = ProjectContext(
+                project_id=stored_project,
+                managed_project_id=creds.managed_project_id if creds else "",
+                tier_id="",
+                source="stored",
+            )
+            return self._project_context
+
+        ctx = resolve_project_context(
+            access_token,
+            configured_project_id=self._configured_project_id,
+            env_project_id=env_project,
+            user_agent_model=model,
+        )
+        # Persist discovered project back to the creds file so the next
+        # session doesn't re-run the discovery.
+        if ctx.project_id or ctx.managed_project_id:
+            google_oauth.update_project_ids(
+                project_id=ctx.project_id,
+                managed_project_id=ctx.managed_project_id,
+            )
+        self._project_context = ctx
+        return ctx
+
+    def _create_chat_completion(
+        self,
+        *,
+        model: str = "gemini-2.5-flash",
+        messages: Optional[List[Dict[str, Any]]] = None,
+        stream: bool = False,
+        tools: Any = None,
+        tool_choice: Any = None,
+        temperature: Optional[float] = None,
+        max_tokens: Optional[int] = None,
+        top_p: Optional[float] = None,
+        stop: Any = None,
+        extra_body: Optional[Dict[str, Any]] = None,
+        timeout: Any = None,
+        **_: Any,
+    ) -> Any:
+        access_token = google_oauth.get_valid_access_token()
+        ctx = self._ensure_project_context(access_token, model)
+
+        thinking_config = None
+        if isinstance(extra_body, dict):
+            thinking_config = extra_body.get("thinking_config") or extra_body.get("thinkingConfig")
+
+        inner = build_gemini_request(
+            messages=messages or [],
+            tools=tools,
+            tool_choice=tool_choice,
+            temperature=temperature,
+            max_tokens=max_tokens,
+            top_p=top_p,
+            stop=stop,
+            thinking_config=thinking_config,
+        )
+        wrapped = wrap_code_assist_request(
+            project_id=ctx.project_id,
+            model=model,
+            inner_request=inner,
+        )
+
+        headers = {
+            "Content-Type": "application/json",
+            "Accept": "application/json",
+            "Authorization": f"Bearer {access_token}",
+            "User-Agent": "hermes-agent (gemini-cli-compat)",
+            "X-Goog-Api-Client": "gl-python/hermes",
+            "x-activity-request-id": str(uuid.uuid4()),
+        }
+        headers.update(self._default_headers)
+
+        if stream:
+            return self._stream_completion(model=model, wrapped=wrapped, headers=headers)
+
+        url = f"{CODE_ASSIST_ENDPOINT}/v1internal:generateContent"
+        response = self._http.post(url, json=wrapped, headers=headers)
+        if response.status_code != 200:
+            raise _gemini_http_error(response)
+        try:
+            payload = response.json()
+        except ValueError as exc:
+            raise CodeAssistError(
+                f"Invalid JSON from Code Assist: {exc}",
+                code="code_assist_invalid_json",
+            ) from exc
+        return _translate_gemini_response(payload, model=model)
+
+    def _stream_completion(
+        self,
+        *,
+        model: str,
+        wrapped: Dict[str, Any],
+        headers: Dict[str, str],
+    ) -> Iterator[_GeminiStreamChunk]:
+        """Generator that yields OpenAI-shaped streaming chunks."""
+        url = f"{CODE_ASSIST_ENDPOINT}/v1internal:streamGenerateContent?alt=sse"
+        stream_headers = dict(headers)
+        stream_headers["Accept"] = "text/event-stream"
+
+        def _generator() -> Iterator[_GeminiStreamChunk]:
+            try:
+                with self._http.stream("POST", url, json=wrapped, headers=stream_headers) as response:
+                    if response.status_code != 200:
+                        # Materialize error body for better diagnostics
+                        response.read()
+                        raise _gemini_http_error(response)
+                    tool_call_counter: List[int] = [0]
+                    for event in _iter_sse_events(response):
+                        for chunk in _translate_stream_event(event, model, tool_call_counter):
+                            yield chunk
+            except httpx.HTTPError as exc:
+                raise CodeAssistError(
+                    f"Streaming request failed: {exc}",
+                    code="code_assist_stream_error",
+                ) from exc
+
+        return _generator()
+
+
+def _gemini_http_error(response: httpx.Response) -> CodeAssistError:
+    """Translate an httpx response into a CodeAssistError with rich metadata.
+
+    Parses Google's error envelope (``{"error": {"code", "message", "status",
+    "details": [...]}}``) so the agent's error classifier can reason about
+    the failure — ``status_code`` enables the rate_limit / auth classification
+    paths, and ``response`` lets the main loop honor ``Retry-After`` just
+    like it does for OpenAI SDK exceptions.
+
+    Also lifts a few recognizable Google conditions into human-readable
+    messages so the user sees something better than a 500-char JSON dump:
+
+        MODEL_CAPACITY_EXHAUSTED → "Gemini model capacity exhausted for
+            <model>. This is a Google-side throttle..."
+        RESOURCE_EXHAUSTED w/o reason → quota-style message
+        404 → "Model <name> not found at cloudcode-pa..."
+    """
+    status = response.status_code
+
+    # Parse the body once, surviving any weird encodings.
+    body_text = ""
+    body_json: Dict[str, Any] = {}
+    try:
+        body_text = response.text
+    except Exception:
+        body_text = ""
+    if body_text:
+        try:
+            parsed = json.loads(body_text)
+            if isinstance(parsed, dict):
+                body_json = parsed
+        except (ValueError, TypeError):
+            body_json = {}
+
+    # Dig into Google's error envelope.  Shape is:
+    #   {"error": {"code": 429, "message": "...", "status": "RESOURCE_EXHAUSTED",
+    #              "details": [{"@type": ".../ErrorInfo", "reason": "MODEL_CAPACITY_EXHAUSTED",
+    #                           "metadata": {...}},
+    #                          {"@type": ".../RetryInfo", "retryDelay": "30s"}]}}
+    err_obj = body_json.get("error") if isinstance(body_json, dict) else None
+    if not isinstance(err_obj, dict):
+        err_obj = {}
+    err_status = str(err_obj.get("status") or "").strip()
+    err_message = str(err_obj.get("message") or "").strip()
+    _raw_details = err_obj.get("details")
+    err_details_list = _raw_details if isinstance(_raw_details, list) else []
+
+    # Extract google.rpc.ErrorInfo reason + metadata.  There may be more
+    # than one ErrorInfo (rare), so we pick the first one with a reason.
+    error_reason = ""
+    error_metadata: Dict[str, Any] = {}
+    retry_delay_seconds: Optional[float] = None
+    for detail in err_details_list:
+        if not isinstance(detail, dict):
+            continue
+        type_url = str(detail.get("@type") or "")
+        if not error_reason and type_url.endswith("/google.rpc.ErrorInfo"):
+            reason = detail.get("reason")
+            if isinstance(reason, str) and reason:
+                error_reason = reason
+            md = detail.get("metadata")
+            if isinstance(md, dict):
+                error_metadata = md
+        elif retry_delay_seconds is None and type_url.endswith("/google.rpc.RetryInfo"):
+            # retryDelay is a google.protobuf.Duration string like "30s" or "1.5s".
+            delay_raw = detail.get("retryDelay")
+            if isinstance(delay_raw, str) and delay_raw.endswith("s"):
+                try:
+                    retry_delay_seconds = float(delay_raw[:-1])
+                except ValueError:
+                    pass
+            elif isinstance(delay_raw, (int, float)):
+                retry_delay_seconds = float(delay_raw)
+
+    # Fall back to the Retry-After header if the body didn't include RetryInfo.
+    if retry_delay_seconds is None:
+        try:
+            header_val = response.headers.get("Retry-After") or response.headers.get("retry-after")
+        except Exception:
+            header_val = None
+        if header_val:
+            try:
+                retry_delay_seconds = float(header_val)
+            except (TypeError, ValueError):
+                retry_delay_seconds = None
+
+    # Classify the error code.  ``code_assist_rate_limited`` stays the default
+    # for 429s; a more specific reason tag helps downstream callers (e.g. tests,
+    # logs) without changing the rate_limit classification path.
+    code = f"code_assist_http_{status}"
+    if status == 401:
+        code = "code_assist_unauthorized"
+    elif status == 429:
+        code = "code_assist_rate_limited"
+        if error_reason == "MODEL_CAPACITY_EXHAUSTED":
+            code = "code_assist_capacity_exhausted"
+
+    # Build a human-readable message.  Keep the status + a raw-body tail for
+    # debugging, but lead with a friendlier summary when we recognize the
+    # Google signal.
+    model_hint = ""
+    if isinstance(error_metadata, dict):
+        model_hint = str(error_metadata.get("model") or error_metadata.get("modelId") or "").strip()
+
+    if status == 429 and error_reason == "MODEL_CAPACITY_EXHAUSTED":
+        target = model_hint or "this Gemini model"
+        message = (
+            f"Gemini capacity exhausted for {target} (Google-side throttle, "
+            f"not a Hermes issue). Try a different Gemini model or set a "
+            f"fallback_providers entry to a non-Gemini provider."
+        )
+        if retry_delay_seconds is not None:
+            message += f" Google suggests retrying in {retry_delay_seconds:g}s."
+    elif status == 429 and err_status == "RESOURCE_EXHAUSTED":
+        message = (
+            f"Gemini quota exhausted ({err_message or 'RESOURCE_EXHAUSTED'}). "
+            f"Check /gquota for remaining daily requests."
+        )
+        if retry_delay_seconds is not None:
+            message += f" Retry suggested in {retry_delay_seconds:g}s."
+    elif status == 404:
+        # Google returns 404 when a model has been retired or renamed.
+        target = model_hint or (err_message or "model")
+        message = (
+            f"Code Assist 404: {target} is not available at "
+            f"cloudcode-pa.googleapis.com. It may have been renamed or "
+            f"retired. Check hermes_cli/models.py for the current list."
+        )
+    elif err_message:
+        # Generic fallback with the parsed message.
+        message = f"Code Assist HTTP {status} ({err_status or 'error'}): {err_message}"
+    else:
+        # Last-ditch fallback — raw body snippet.
+        message = f"Code Assist returned HTTP {status}: {body_text[:500]}"
+
+    return CodeAssistError(
+        message,
+        code=code,
+        status_code=status,
+        response=response,
+        retry_after=retry_delay_seconds,
+        details={
+            "status": err_status,
+            "reason": error_reason,
+            "metadata": error_metadata,
+            "message": err_message,
+        },
+    )

agent/gemini_native_adapter.py

@@ -27,18 +27,10 @@ from typing import Any, Dict, Iterator, List, Optional
 
 import httpx
 
-from agent.bounded_response import read_streaming_error_body
 from agent.gemini_schema import sanitize_gemini_tool_parameters
 
 logger = logging.getLogger(__name__)
 
-try:
-    import hermes_cli as _hermes_cli
-
-    _HERMES_VERSION = str(_hermes_cli.__version__)
-except Exception:
-    _HERMES_VERSION = "0.0.0"
-
 DEFAULT_GEMINI_BASE_URL = "https://generativelanguage.googleapis.com/v1beta"
 
 # Published max output-token ceiling shared by every current Gemini text model
@@ -106,10 +98,7 @@ def probe_gemini_tier(
                 url,
                 params={"key": key},
                 json=payload,
-                headers={
-                    "Content-Type": "application/json",
-                    "X-Goog-Api-Client": f"hermes-agent/{_HERMES_VERSION}",
-                },
+                headers={"Content-Type": "application/json"},
             )
     except Exception as exc:
         logger.debug("probe_gemini_tier: network error: %s", exc)
@@ -270,12 +259,8 @@ def _translate_tool_call_to_gemini(tool_call: Dict[str, Any]) -> Dict[str, Any]:
         }
     }
     thought_signature = _tool_call_extra_signature(tool_call)
-    # Fallback sentinel for cross-provider tool_calls (e.g. fallback from
-    # xAI/Anthropic to Gemini, where the original tool_call carries no
-    # Gemini thoughtSignature). Mirrors gemini_cloudcode_adapter.py:106.
-    # Without this, Gemini 3 thinking models reject replayed history with
-    # 400 INVALID_ARGUMENT on the missing thoughtSignature.
-    part["thoughtSignature"] = thought_signature or "skip_thought_signature_validator"
+    if thought_signature:
+        part["thoughtSignature"] = thought_signature
     return part
 
 
@@ -352,22 +337,6 @@ def _build_gemini_contents(messages: List[Dict[str, Any]]) -> tuple[List[Dict[st
         if parts:
             contents.append({"role": gemini_role, "parts": parts})
 
-    # Gemini's generateContent requires strict user/model alternation;
-    # consecutive same-role contents are rejected with HTTP 400 "Please ensure
-    # that multiturn requests alternate between user and model". The loop above
-    # emits one content per source message, so parallel tool calls (N tool
-    # results become N user functionResponse contents), back-to-back user turns,
-    # or merged assistant turns would each violate that. Merge adjacent
-    # same-role contents by concatenating their parts. For parallel calls this
-    # also produces the grouped multi-functionResponse turn Gemini expects.
-    merged_contents: List[Dict[str, Any]] = []
-    for content in contents:
-        if merged_contents and merged_contents[-1]["role"] == content["role"]:
-            merged_contents[-1]["parts"].extend(content["parts"])
-        else:
-            merged_contents.append(content)
-    contents = merged_contents
-
     system_instruction = None
     joined_system = "\n".join(part for part in system_text_parts if part).strip()
     if joined_system:
@@ -757,17 +726,14 @@ def translate_stream_event(event: Dict[str, Any], model: str, tool_call_indices:
     return chunks
 
 
-def gemini_http_error(
-    response: httpx.Response, *, body_text: Optional[str] = None
-) -> GeminiAPIError:
+def gemini_http_error(response: httpx.Response) -> GeminiAPIError:
     status = response.status_code
+    body_text = ""
     body_json: Dict[str, Any] = {}
-    if body_text is None:
-        try:
-            body_text = response.text
-        except Exception:
-            body_text = ""
-    body_text = body_text or ""
+    try:
+        body_text = response.text
+    except Exception:
+        body_text = ""
     if body_text:
         try:
             parsed = json.loads(body_text)
@@ -915,11 +881,7 @@ class GeminiNativeClient:
             "Content-Type": "application/json",
             "Accept": "application/json",
             "x-goog-api-key": self.api_key,
-            # Include Hermes client context following Gemini's partner
-            # integration guidance.
-            # See https://ai.google.dev/gemini-api/docs/partner-integration
-            "User-Agent": f"hermes-agent/{_HERMES_VERSION} (gemini-native)",
-            "X-Goog-Api-Client": f"hermes-agent/{_HERMES_VERSION}",
+            "User-Agent": "hermes-agent (gemini-native)",
         }
         headers.update(self._default_headers)
         return headers
@@ -990,8 +952,8 @@ class GeminiNativeClient:
             try:
                 with self._http.stream("POST", url, json=request, headers=stream_headers, timeout=timeout) as response:
                     if response.status_code != 200:
-                        body_text = read_streaming_error_body(response)
-                        raise gemini_http_error(response, body_text=body_text)
+                        response.read()
+                        raise gemini_http_error(response)
                     tool_call_indices: Dict[str, Dict[str, Any]] = {}
                     for event in _iter_sse_events(response):
                         for chunk in translate_stream_event(event, model, tool_call_indices):

agent/gemini_schema.py

@@ -87,30 +87,6 @@ def sanitize_gemini_schema(schema: Any) -> Dict[str, Any]:
         if any(not isinstance(item, str) for item in enum_val):
             cleaned.pop("enum", None)
 
-    # Gemini validates ``required`` strictly against the same node's
-    # ``properties`` — GenerateContentRequest fails with HTTP 400
-    # "...items.required[0]: property is not defined" when a required name
-    # has no matching property in that node.  MCP servers routinely emit
-    # this shape (e.g. the GitHub remote MCP's array item schemas carry
-    # ``required`` without ``properties``), and one bad tool schema fails
-    # the ENTIRE request before any model output.  Filter ``required`` to
-    # names that exist in this node's ``properties`` and drop it when
-    # nothing valid remains.  The tool handler still validates required
-    # fields at execution time, so this only removes what Gemini couldn't
-    # accept anyway.  (Port of Kilo-Org/kilocode#11955.)
-    required_val = cleaned.get("required")
-    if isinstance(required_val, list):
-        props_val = cleaned.get("properties")
-        prop_names = set(props_val.keys()) if isinstance(props_val, dict) else set()
-        valid_required = [
-            name for name in required_val
-            if isinstance(name, str) and name in prop_names
-        ]
-        if not valid_required:
-            cleaned.pop("required", None)
-        elif len(valid_required) != len(required_val):
-            cleaned["required"] = valid_required
-
     return cleaned
 
 

agent/google_code_assist.py

@@ -0,0 +1,451 @@
+"""Google Code Assist API client — project discovery, onboarding, quota.
+
+The Code Assist API powers Google's official gemini-cli. It sits at
+``cloudcode-pa.googleapis.com`` and provides:
+
+- Free tier access (generous daily quota) for personal Google accounts
+- Paid tier access via GCP projects with billing / Workspace / Standard / Enterprise
+
+This module handles the control-plane dance needed before inference:
+
+1. ``load_code_assist()`` — probe the user's account to learn what tier they're on
+   and whether a ``cloudaicompanionProject`` is already assigned.
+2. ``onboard_user()`` — if the user hasn't been onboarded yet (new account, fresh
+   free tier, etc.), call this with the chosen tier + project id. Supports LRO
+   polling for slow provisioning.
+3. ``retrieve_user_quota()`` — fetch the ``buckets[]`` array showing remaining
+   quota per model, used by the ``/gquota`` slash command.
+
+VPC-SC handling: enterprise accounts under a VPC Service Controls perimeter
+will get ``SECURITY_POLICY_VIOLATED`` on ``load_code_assist``. We catch this
+and force the account to ``standard-tier`` so the call chain still succeeds.
+
+Derived from opencode-gemini-auth (MIT) and clawdbot/extensions/google. The
+request/response shapes are specific to Google's internal Code Assist API,
+documented nowhere public — we copy them from the reference implementations.
+"""
+
+from __future__ import annotations
+
+import json
+import logging
+import time
+import urllib.error
+import urllib.request
+import uuid
+from dataclasses import dataclass, field
+from typing import Any, Dict, List, Optional
+
+logger = logging.getLogger(__name__)
+
+
+# =============================================================================
+# Constants
+# =============================================================================
+
+CODE_ASSIST_ENDPOINT = "https://cloudcode-pa.googleapis.com"
+
+# Fallback endpoints tried when prod returns an error during project discovery
+FALLBACK_ENDPOINTS = [
+    "https://daily-cloudcode-pa.sandbox.googleapis.com",
+    "https://autopush-cloudcode-pa.sandbox.googleapis.com",
+]
+
+# Tier identifiers that Google's API uses
+FREE_TIER_ID = "free-tier"
+LEGACY_TIER_ID = "legacy-tier"
+STANDARD_TIER_ID = "standard-tier"
+
+# Default HTTP headers matching gemini-cli's fingerprint.
+# Google may reject unrecognized User-Agents on these internal endpoints.
+_GEMINI_CLI_USER_AGENT = "google-api-nodejs-client/9.15.1 (gzip)"
+_X_GOOG_API_CLIENT = "gl-node/24.0.0"
+_DEFAULT_REQUEST_TIMEOUT = 30.0
+_ONBOARDING_POLL_ATTEMPTS = 12
+_ONBOARDING_POLL_INTERVAL_SECONDS = 5.0
+
+
+class CodeAssistError(RuntimeError):
+    """Exception raised by the Code Assist (``cloudcode-pa``) integration.
+
+    Carries HTTP status / response / retry-after metadata so the agent's
+    ``error_classifier._extract_status_code`` and the main loop's Retry-After
+    handling (which walks ``error.response.headers``) pick up the right
+    signals.  Without these, 429s from the OAuth path look like opaque
+    ``RuntimeError`` and skip the rate-limit path.
+    """
+
+    def __init__(
+        self,
+        message: str,
+        *,
+        code: str = "code_assist_error",
+        status_code: Optional[int] = None,
+        response: Any = None,
+        retry_after: Optional[float] = None,
+        details: Optional[Dict[str, Any]] = None,
+    ) -> None:
+        super().__init__(message)
+        self.code = code
+        # ``status_code`` is picked up by ``agent.error_classifier._extract_status_code``
+        # so a 429 from Code Assist classifies as FailoverReason.rate_limit and
+        # triggers the main loop's fallback_providers chain the same way SDK
+        # errors do.
+        self.status_code = status_code
+        # ``response`` is the underlying ``httpx.Response`` (or a shim with a
+        # ``.headers`` mapping and ``.json()`` method).  The main loop reads
+        # ``error.response.headers["Retry-After"]`` to honor Google's retry
+        # hints when the backend throttles us.
+        self.response = response
+        # Parsed ``Retry-After`` seconds (kept separately for convenience —
+        # Google returns retry hints in both the header and the error body's
+        # ``google.rpc.RetryInfo`` details, and we pick whichever we found).
+        self.retry_after = retry_after
+        # Parsed structured error details from the Google error envelope
+        # (e.g. ``{"reason": "MODEL_CAPACITY_EXHAUSTED", "status": "RESOURCE_EXHAUSTED"}``).
+        # Useful for logging and for tests that want to assert on specifics.
+        self.details = details or {}
+
+
+class ProjectIdRequiredError(CodeAssistError):
+    def __init__(self, message: str = "GCP project id required for this tier") -> None:
+        super().__init__(message, code="code_assist_project_id_required")
+
+
+# =============================================================================
+# HTTP primitive (auth via Bearer token passed per-call)
+# =============================================================================
+
+def _build_headers(access_token: str, *, user_agent_model: str = "") -> Dict[str, str]:
+    ua = _GEMINI_CLI_USER_AGENT
+    if user_agent_model:
+        ua = f"{ua} model/{user_agent_model}"
+    return {
+        "Content-Type": "application/json",
+        "Accept": "application/json",
+        "Authorization": f"Bearer {access_token}",
+        "User-Agent": ua,
+        "X-Goog-Api-Client": _X_GOOG_API_CLIENT,
+        "x-activity-request-id": str(uuid.uuid4()),
+    }
+
+
+def _client_metadata() -> Dict[str, str]:
+    """Match Google's gemini-cli exactly — unrecognized metadata may be rejected."""
+    return {
+        "ideType": "IDE_UNSPECIFIED",
+        "platform": "PLATFORM_UNSPECIFIED",
+        "pluginType": "GEMINI",
+    }
+
+
+def _post_json(
+    url: str,
+    body: Dict[str, Any],
+    access_token: str,
+    *,
+    timeout: float = _DEFAULT_REQUEST_TIMEOUT,
+    user_agent_model: str = "",
+) -> Dict[str, Any]:
+    data = json.dumps(body).encode("utf-8")
+    request = urllib.request.Request(
+        url, data=data, method="POST",
+        headers=_build_headers(access_token, user_agent_model=user_agent_model),
+    )
+    try:
+        with urllib.request.urlopen(request, timeout=timeout) as response:
+            raw = response.read().decode("utf-8", errors="replace")
+            return json.loads(raw) if raw else {}
+    except urllib.error.HTTPError as exc:
+        detail = ""
+        try:
+            detail = exc.read().decode("utf-8", errors="replace")
+        except Exception:
+            pass
+        # Special case: VPC-SC violation should be distinguishable
+        if _is_vpc_sc_violation(detail):
+            raise CodeAssistError(
+                f"VPC-SC policy violation: {detail}",
+                code="code_assist_vpc_sc",
+            ) from exc
+        raise CodeAssistError(
+            f"Code Assist HTTP {exc.code}: {detail or exc.reason}",
+            code=f"code_assist_http_{exc.code}",
+        ) from exc
+    except urllib.error.URLError as exc:
+        raise CodeAssistError(
+            f"Code Assist request failed: {exc}",
+            code="code_assist_network_error",
+        ) from exc
+
+
+def _is_vpc_sc_violation(body: str) -> bool:
+    """Detect a VPC Service Controls violation from a response body."""
+    if not body:
+        return False
+    try:
+        parsed = json.loads(body)
+    except (json.JSONDecodeError, ValueError):
+        return "SECURITY_POLICY_VIOLATED" in body
+    # Walk the nested error structure Google uses
+    error = parsed.get("error") if isinstance(parsed, dict) else None
+    if not isinstance(error, dict):
+        return False
+    details = error.get("details") or []
+    if isinstance(details, list):
+        for item in details:
+            if isinstance(item, dict):
+                reason = item.get("reason") or ""
+                if reason == "SECURITY_POLICY_VIOLATED":
+                    return True
+    msg = str(error.get("message", ""))
+    return "SECURITY_POLICY_VIOLATED" in msg
+
+
+# =============================================================================
+# load_code_assist — discovers current tier + assigned project
+# =============================================================================
+
+@dataclass
+class CodeAssistProjectInfo:
+    """Result from ``load_code_assist``."""
+    current_tier_id: str = ""
+    cloudaicompanion_project: str = ""   # Google-managed project (free tier)
+    allowed_tiers: List[str] = field(default_factory=list)
+    raw: Dict[str, Any] = field(default_factory=dict)
+
+
+def load_code_assist(
+    access_token: str,
+    *,
+    project_id: str = "",
+    user_agent_model: str = "",
+) -> CodeAssistProjectInfo:
+    """Call ``POST /v1internal:loadCodeAssist`` with prod → sandbox fallback.
+
+    Returns whatever tier + project info Google reports. On VPC-SC violations,
+    returns a synthetic ``standard-tier`` result so the chain can continue.
+    """
+    body: Dict[str, Any] = {
+        "metadata": {
+            "duetProject": project_id,
+            **_client_metadata(),
+        },
+    }
+    if project_id:
+        body["cloudaicompanionProject"] = project_id
+
+    endpoints = [CODE_ASSIST_ENDPOINT] + FALLBACK_ENDPOINTS
+    last_err: Optional[Exception] = None
+    for endpoint in endpoints:
+        url = f"{endpoint}/v1internal:loadCodeAssist"
+        try:
+            resp = _post_json(url, body, access_token, user_agent_model=user_agent_model)
+            return _parse_load_response(resp)
+        except CodeAssistError as exc:
+            if exc.code == "code_assist_vpc_sc":
+                logger.info("VPC-SC violation on %s — defaulting to standard-tier", endpoint)
+                return CodeAssistProjectInfo(
+                    current_tier_id=STANDARD_TIER_ID,
+                    cloudaicompanion_project=project_id,
+                )
+            last_err = exc
+            logger.warning("loadCodeAssist failed on %s: %s", endpoint, exc)
+            continue
+    if last_err:
+        raise last_err
+    return CodeAssistProjectInfo()
+
+
+def _parse_load_response(resp: Dict[str, Any]) -> CodeAssistProjectInfo:
+    current_tier = resp.get("currentTier") or {}
+    tier_id = str(current_tier.get("id") or "") if isinstance(current_tier, dict) else ""
+    project = str(resp.get("cloudaicompanionProject") or "")
+    allowed = resp.get("allowedTiers") or []
+    allowed_ids: List[str] = []
+    if isinstance(allowed, list):
+        for t in allowed:
+            if isinstance(t, dict):
+                tid = str(t.get("id") or "")
+                if tid:
+                    allowed_ids.append(tid)
+    return CodeAssistProjectInfo(
+        current_tier_id=tier_id,
+        cloudaicompanion_project=project,
+        allowed_tiers=allowed_ids,
+        raw=resp,
+    )
+
+
+# =============================================================================
+# onboard_user — provisions a new user on a tier (with LRO polling)
+# =============================================================================
+
+def onboard_user(
+    access_token: str,
+    *,
+    tier_id: str,
+    project_id: str = "",
+    user_agent_model: str = "",
+) -> Dict[str, Any]:
+    """Call ``POST /v1internal:onboardUser`` to provision the user.
+
+    For paid tiers, ``project_id`` is REQUIRED (raises ProjectIdRequiredError).
+    For free tiers, ``project_id`` is optional — Google will assign one.
+
+    Returns the final operation response. Polls ``/v1internal/<name>`` for up
+    to ``_ONBOARDING_POLL_ATTEMPTS`` × ``_ONBOARDING_POLL_INTERVAL_SECONDS``
+    (default: 12 × 5s = 1 min).
+    """
+    if tier_id != FREE_TIER_ID and tier_id != LEGACY_TIER_ID and not project_id:
+        raise ProjectIdRequiredError(
+            f"Tier {tier_id!r} requires a GCP project id. "
+            "Set HERMES_GEMINI_PROJECT_ID or GOOGLE_CLOUD_PROJECT."
+        )
+
+    body: Dict[str, Any] = {
+        "tierId": tier_id,
+        "metadata": _client_metadata(),
+    }
+    if project_id:
+        body["cloudaicompanionProject"] = project_id
+
+    endpoint = CODE_ASSIST_ENDPOINT
+    url = f"{endpoint}/v1internal:onboardUser"
+    resp = _post_json(url, body, access_token, user_agent_model=user_agent_model)
+
+    # Poll if LRO (long-running operation)
+    if not resp.get("done"):
+        op_name = resp.get("name", "")
+        if not op_name:
+            return resp
+        for attempt in range(_ONBOARDING_POLL_ATTEMPTS):
+            time.sleep(_ONBOARDING_POLL_INTERVAL_SECONDS)
+            poll_url = f"{endpoint}/v1internal/{op_name}"
+            try:
+                poll_resp = _post_json(poll_url, {}, access_token, user_agent_model=user_agent_model)
+            except CodeAssistError as exc:
+                logger.warning("Onboarding poll attempt %d failed: %s", attempt + 1, exc)
+                continue
+            if poll_resp.get("done"):
+                return poll_resp
+        logger.warning("Onboarding did not complete within %d attempts", _ONBOARDING_POLL_ATTEMPTS)
+    return resp
+
+
+# =============================================================================
+# retrieve_user_quota — for /gquota
+# =============================================================================
+
+@dataclass
+class QuotaBucket:
+    model_id: str
+    token_type: str = ""
+    remaining_fraction: float = 0.0
+    reset_time_iso: str = ""
+    raw: Dict[str, Any] = field(default_factory=dict)
+
+
+def retrieve_user_quota(
+    access_token: str,
+    *,
+    project_id: str = "",
+    user_agent_model: str = "",
+) -> List[QuotaBucket]:
+    """Call ``POST /v1internal:retrieveUserQuota`` and parse ``buckets[]``."""
+    body: Dict[str, Any] = {}
+    if project_id:
+        body["project"] = project_id
+    url = f"{CODE_ASSIST_ENDPOINT}/v1internal:retrieveUserQuota"
+    resp = _post_json(url, body, access_token, user_agent_model=user_agent_model)
+    raw_buckets = resp.get("buckets") or []
+    buckets: List[QuotaBucket] = []
+    if not isinstance(raw_buckets, list):
+        return buckets
+    for b in raw_buckets:
+        if not isinstance(b, dict):
+            continue
+        buckets.append(QuotaBucket(
+            model_id=str(b.get("modelId") or ""),
+            token_type=str(b.get("tokenType") or ""),
+            remaining_fraction=float(b.get("remainingFraction") or 0.0),
+            reset_time_iso=str(b.get("resetTime") or ""),
+            raw=b,
+        ))
+    return buckets
+
+
+# =============================================================================
+# Project context resolution
+# =============================================================================
+
+@dataclass
+class ProjectContext:
+    """Resolved state for a given OAuth session."""
+    project_id: str = ""           # effective project id sent on requests
+    managed_project_id: str = ""   # Google-assigned project (free tier)
+    tier_id: str = ""
+    source: str = ""               # "env", "config", "discovered", "onboarded"
+
+
+def resolve_project_context(
+    access_token: str,
+    *,
+    configured_project_id: str = "",
+    env_project_id: str = "",
+    user_agent_model: str = "",
+) -> ProjectContext:
+    """Figure out what project id + tier to use for requests.
+
+    Priority:
+      1. If configured_project_id or env_project_id is set, use that directly
+         and short-circuit (no discovery needed).
+      2. Otherwise call loadCodeAssist to see what Google says.
+      3. If no tier assigned yet, onboard the user (free tier default).
+    """
+    # Short-circuit: caller provided a project id
+    if configured_project_id:
+        return ProjectContext(
+            project_id=configured_project_id,
+            tier_id=STANDARD_TIER_ID,  # assume paid since they specified one
+            source="config",
+        )
+    if env_project_id:
+        return ProjectContext(
+            project_id=env_project_id,
+            tier_id=STANDARD_TIER_ID,
+            source="env",
+        )
+
+    # Discover via loadCodeAssist
+    info = load_code_assist(access_token, user_agent_model=user_agent_model)
+
+    effective_project = info.cloudaicompanion_project
+    tier = info.current_tier_id
+
+    if not tier:
+        # User hasn't been onboarded — provision them on free tier
+        onboard_resp = onboard_user(
+            access_token,
+            tier_id=FREE_TIER_ID,
+            project_id="",
+            user_agent_model=user_agent_model,
+        )
+        # Re-parse from the onboard response
+        response_body = onboard_resp.get("response") or {}
+        if isinstance(response_body, dict):
+            effective_project = (
+                effective_project
+                or str(response_body.get("cloudaicompanionProject") or "")
+            )
+        tier = FREE_TIER_ID
+        source = "onboarded"
+    else:
+        source = "discovered"
+
+    return ProjectContext(
+        project_id=effective_project,
+        managed_project_id=effective_project if tier == FREE_TIER_ID else "",
+        tier_id=tier,
+        source=source,
+    )

agent/i18n.py

@@ -25,14 +25,14 @@ Language resolution order:
     3. ``display.language`` from config.yaml
     4. ``"en"`` (baseline)
 
-Supported languages: en, zh, zh-hant, ja, de, es, fr, tr, uk, af, ko, it, ga,
-pt, ru, hu, ar.  Unknown values fall back to en.
+Supported languages: en, zh, ja, de, es, fr, tr, uk.  Unknown values fall back to en.
 """
 
 from __future__ import annotations
 
 import logging
 import os
+import sysconfig
 import threading
 from functools import lru_cache
 from pathlib import Path
@@ -42,7 +42,7 @@ logger = logging.getLogger(__name__)
 
 SUPPORTED_LANGUAGES: tuple[str, ...] = (
     "en", "zh", "zh-hant", "ja", "de", "es", "fr", "tr", "uk",
-    "af", "ko", "it", "ga", "pt", "ru", "hu", "ar",
+    "af", "ko", "it", "ga", "pt", "ru", "hu",
 )
 DEFAULT_LANGUAGE = "en"
 
@@ -79,9 +79,6 @@ _LANGUAGE_ALIASES: dict[str, str] = {
     "russian": "ru", "русский": "ru", "ru-ru": "ru",
     # Hungarian
     "hungarian": "hu", "magyar": "hu", "hu-hu": "hu",
-    # Arabic — bare "arabic"/endonym plus the common regional BCP-47 tags.
-    "arabic": "ar", "العربية": "ar",
-    "ar-sa": "ar", "ar-eg": "ar", "ar-ae": "ar", "ar-ma": "ar", "ar-dz": "ar",
 }
 
 _catalog_cache: dict[str, dict[str, str]] = {}
@@ -95,8 +92,12 @@ def _locales_dir() -> Path:
 
     1. ``HERMES_BUNDLED_LOCALES`` env var -- set by the Nix wrapper (or any
        sealed-packaging system) to point at the installed catalog directory.
-    2. ``<repo-root>/locales`` -- source checkouts and editable installs,
+    2. ``<repo-root>/locales`` -- source checkouts and ``pip install -e .``,
        where the working tree sits next to ``agent/``.
+    3. ``<sysconfig data|purelib|platlib>/locales`` -- pip wheel installs.
+       setuptools ``data-files`` extracts ``locales/*.yaml`` under the
+       interpreter's ``data`` scheme; the other schemes are checked as a
+       safety net for nonstandard layouts.
 
     Falling through to the source-style path (even when missing) keeps
     ``_load_catalog`` error messages informative -- it logs the path it
@@ -115,6 +116,25 @@ def _locales_dir() -> Path:
 
     # agent/i18n.py -> agent/ -> repo root (source checkout, editable install)
     source_dir = Path(__file__).resolve().parent.parent / "locales"
+    if source_dir.is_dir():
+        return source_dir
+
+    # pip wheel install: data-files lands under the interpreter data scheme.
+    # ``data`` (== sys.prefix in a venv) is where setuptools data-files extract
+    # and is checked first. ``purelib``/``platlib`` (site-packages) are a safety
+    # net for nonstandard layouts. NOTE: this does NOT cover ``pip install
+    # --user`` (user scheme, ~/.local/locales) or ``pip install --target`` --
+    # both are out of scope; see the plan header.
+    for scheme in ("data", "purelib", "platlib"):
+        raw = sysconfig.get_path(scheme)
+        if not raw:
+            continue
+        candidate = Path(raw) / "locales"
+        if candidate.is_dir():
+            return candidate
+
+    # Last resort: return the source-style path so _load_catalog's catalog-missing
+    # log (logger.debug "i18n catalog missing for %s at %s") stays informative.
     return source_dir
 
 

agent/image_routing.py

@@ -17,17 +17,13 @@ It reads ``agent.image_input_mode`` from config.yaml (``auto`` | ``native``
 | ``text``, default ``auto``) and the active model's capability metadata.
 
 In ``auto`` mode:
-  - If the active model reports ``supports_vision=True`` (via config
-    override or models.dev metadata), we attach natively — vision-capable
-    main models should always see the original pixels, even when an
-    auxiliary vision backend is configured. That auxiliary backend then
-    acts as a *fallback* for sessions whose main model can't take images.
-  - Otherwise, if the user has explicitly configured ``auxiliary.vision``
-    (provider/model/base_url not ``auto``/empty), we route through the
-    text pipeline so the auxiliary vision backend can describe the image
-    for the text-only main model.
-  - Otherwise (non-vision model, no explicit override), we fall back to
-    text via the default vision_analyze flow.
+  - If the user has explicitly configured ``auxiliary.vision.provider``
+    (i.e. not ``auto`` and not empty), we assume they want the text pipeline
+    regardless of the main model — they've opted in to a specific vision
+    backend for a reason (cost, quality, local-only, etc.).
+  - Otherwise, if the active model reports ``supports_vision=True`` in its
+    models.dev metadata, we attach natively.
+  - Otherwise (non-vision model, no explicit override), we fall back to text.
 
 This keeps ``vision_analyze`` surfaced as a tool in every session — skills
 and agent flows that chain it (browser screenshots, deeper inspection of
@@ -181,8 +177,6 @@ def _supports_vision_override(
     cfg: Optional[Dict[str, Any]],
     provider: str,
     model: str,
-    *,
-    requested_provider: str = "",
 ) -> Optional[bool]:
     """Resolve user-declared vision capability from config.yaml.
 
@@ -190,14 +184,8 @@ def _supports_vision_override(
       1. ``model.supports_vision`` (top-level shortcut for the active model)
       2. ``providers.<provider>.models.<model>.supports_vision``
          (named custom providers — ``provider`` may be the runtime-resolved
-         value ``"custom"``, the runtime's originally requested provider,
-         and/or the user-declared name under ``model.provider``; all are
-         tried. For ``custom:<name>`` syntax, the stripped ``<name>`` is also
-         tried as a provider key.)
-      2b. ``custom_providers`` (legacy list form) ``.models.<model>``
-
-    Under (2) and (2b), the per-model capability key may be written as
-    either ``supports_vision`` or the shorter ``vision`` alias; both work.
+         value ``"custom"`` and/or the user-declared name under
+         ``model.provider``; both are tried)
 
     Returns None when no override is set, so the caller falls through to
     models.dev. Returns False explicitly only when the user wrote a
@@ -217,30 +205,18 @@ def _supports_vision_override(
     # get rewritten to provider="custom" at runtime
     # (hermes_cli/runtime_provider.py:_resolve_named_custom_runtime), so the
     # config still holds the user-declared name under model.provider. Try
-    # both as candidate provider keys. Either identity may use the
-    # "custom:<name>" form while providers: is keyed by bare <name>.
+    # both as candidate provider keys.
     config_provider = str(model_cfg.get("provider") or "").strip()
-    provider_candidates: List[str] = []
-    for candidate in (requested_provider, provider, config_provider):
-        if not candidate:
-            continue
-        provider_candidates.append(candidate)
-        if candidate.startswith("custom:"):
-            stripped_candidate = candidate[len("custom:"):]
-            if stripped_candidate:
-                provider_candidates.append(stripped_candidate)
     providers_raw = cfg.get("providers")
     providers_cfg: Dict[str, Any] = providers_raw if isinstance(providers_raw, dict) else {}
-    for p in dict.fromkeys(provider_candidates):
+    for p in dict.fromkeys(filter(None, (provider, config_provider))):
         entry_raw = providers_cfg.get(p)
         entry: Dict[str, Any] = entry_raw if isinstance(entry_raw, dict) else {}
         models_raw = entry.get("models")
         models_cfg: Dict[str, Any] = models_raw if isinstance(models_raw, dict) else {}
         per_model_raw = models_cfg.get(model)
         per_model: Dict[str, Any] = per_model_raw if isinstance(per_model_raw, dict) else {}
-        coerced = _coerce_capability_bool(
-            per_model.get("supports_vision", per_model.get("vision"))
-        )
+        coerced = _coerce_capability_bool(per_model.get("supports_vision"))
         if coerced is not None:
             return coerced
 
@@ -249,102 +225,30 @@ def _supports_vision_override(
     # may appear as the raw name or "custom:<name>" at runtime).
     custom_providers = cfg.get("custom_providers")
     if isinstance(custom_providers, list):
-        # Candidate priority matters when the CLI-selected provider differs
-        # from model.provider. Walk identities first, then config entries, so
-        # list order cannot let the persisted default shadow the live route.
-        for candidate in dict.fromkeys(provider_candidates):
-            candidate_name = candidate.strip().lower()
-            for entry_raw in custom_providers:
-                if not isinstance(entry_raw, dict):
-                    continue
-                entry_name = str(entry_raw.get("name") or "").strip().lower()
-                if entry_name != candidate_name:
-                    continue
-                models_raw = entry_raw.get("models")
-                models_cfg = models_raw if isinstance(models_raw, dict) else {}
-                per_model_raw = models_cfg.get(model)
-                per_model = per_model_raw if isinstance(per_model_raw, dict) else {}
-                coerced = _coerce_capability_bool(
-                    per_model.get("supports_vision", per_model.get("vision"))
-                )
-                if coerced is not None:
-                    return coerced
-
-    return None
-
-
-def _resolve_inference_base_url(
-    cfg: Optional[Dict[str, Any]],
-    provider: str,
-) -> str:
-    """Best-effort base URL for the active inference provider."""
-    try:
-        from agent.auxiliary_client import _runtime_main_value
-
-        runtime = str(_runtime_main_value("base_url") or "").strip()
-        runtime_provider = str(_runtime_main_value("provider") or "").strip().lower()
-        requested_provider = str(provider or "").strip().lower()
-        if runtime and (not requested_provider or requested_provider == runtime_provider):
-            return runtime
-    except Exception:
-        pass
-
-    if not isinstance(cfg, dict):
-        return ""
-
-    model_cfg_raw = cfg.get("model")
-    model_cfg: Dict[str, Any] = model_cfg_raw if isinstance(model_cfg_raw, dict) else {}
-    base_url = str(model_cfg.get("base_url") or "").strip()
-    if base_url:
-        return base_url
-
-    config_provider = str(model_cfg.get("provider") or "").strip()
-    candidate_names: set[str] = set()
-    for p in filter(None, (provider, config_provider)):
-        candidate_names.add(p)
-        if p.lower().startswith("custom:"):
-            candidate_names.add(p.split(":", 1)[1])
-        else:
-            candidate_names.add(f"custom:{p}")
-
-    providers_cfg = cfg.get("providers")
-    if isinstance(providers_cfg, dict):
-        for name in candidate_names:
-            entry = providers_cfg.get(name)
-            if isinstance(entry, dict):
-                bu = str(entry.get("base_url") or "").strip()
-                if bu:
-                    return bu
-
-    custom_providers = cfg.get("custom_providers")
-    if isinstance(custom_providers, list):
-        lowered = {n.lower() for n in candidate_names}
+        # Build candidate names: the provider value and the config provider
+        # value, both raw and with "custom:" prefix stripped/added.
+        candidate_names: set = set()
+        for p in filter(None, (provider, config_provider)):
+            candidate_names.add(p)
+            if p.startswith("custom:"):
+                candidate_names.add(p[len("custom:"):])
+            else:
+                candidate_names.add(f"custom:{p}")
         for entry_raw in custom_providers:
             if not isinstance(entry_raw, dict):
                 continue
             entry_name = str(entry_raw.get("name") or "").strip()
-            if entry_name not in candidate_names and entry_name.lower() not in lowered:
+            if entry_name not in candidate_names:
                 continue
-            bu = str(entry_raw.get("base_url") or "").strip()
-            if bu:
-                return bu
+            models_raw = entry_raw.get("models")
+            models_cfg = models_raw if isinstance(models_raw, dict) else {}
+            per_model_raw = models_cfg.get(model)
+            per_model = per_model_raw if isinstance(per_model_raw, dict) else {}
+            coerced = _coerce_capability_bool(per_model.get("supports_vision"))
+            if coerced is not None:
+                return coerced
 
-    return ""
-
-
-def _should_probe_ollama_vision(provider: str, base_url: str) -> bool:
-    """True when the active provider likely fronts a local Ollama server."""
-    p = (provider or "").strip().lower()
-    if p == "ollama":
-        return True
-    if not base_url:
-        return False
-    try:
-        from agent.model_metadata import detect_local_server_type
-
-        return detect_local_server_type(base_url) == "ollama"
-    except Exception:
-        return False
+    return None
 
 
 def _coerce_mode(raw: Any) -> str:
@@ -360,10 +264,8 @@ def _coerce_mode(raw: Any) -> str:
 def _explicit_aux_vision_override(cfg: Optional[Dict[str, Any]]) -> bool:
     """True when the user configured a specific auxiliary vision backend.
 
-    An explicit override means the user has a dedicated vision backend
-    available; it's used as a *fallback* when the main model can't take
-    images natively. In ``auto`` mode, native vision on a vision-capable
-    main model still wins over this fallback — see issue #29135.
+    An explicit override means the user *wants* the text pipeline (they're
+    paying for a dedicated vision model), so we don't silently bypass it.
     """
     if not isinstance(cfg, dict):
         return False
@@ -388,8 +290,6 @@ def _lookup_supports_vision(
     provider: str,
     model: str,
     cfg: Optional[Dict[str, Any]] = None,
-    *,
-    requested_provider: str = "",
 ) -> Optional[bool]:
     """Return True/False if we can resolve caps, None if unknown.
 
@@ -397,73 +297,26 @@ def _lookup_supports_vision(
     (so custom/local models declared as vision-capable don't fall through to
     text routing in ``auto`` mode), then falls back to models.dev.
     """
-    # Named custom providers are canonicalized to ``provider="custom"`` by
-    # runtime resolution.  The original CLI/config name is carried in the
-    # context-local main runtime so capability lookup can still select the
-    # exact custom_providers entry.  Require an exact provider+model match:
-    # background/auxiliary lookups must never borrow another turn's identity.
-    if not requested_provider:
-        try:
-            from agent.auxiliary_client import _runtime_main_value
-
-            runtime_provider = str(
-                _runtime_main_value("provider") or ""
-            ).strip().lower()
-            runtime_model = str(_runtime_main_value("model") or "").strip()
-            lookup_provider = str(provider or "").strip().lower()
-            lookup_model = str(model or "").strip()
-            if runtime_provider == lookup_provider and runtime_model == lookup_model:
-                requested_provider = str(
-                    _runtime_main_value("requested_provider") or ""
-                ).strip()
-        except Exception:
-            pass
-
-    override = _supports_vision_override(
-        cfg,
-        provider,
-        model,
-        requested_provider=requested_provider,
-    )
+    override = _supports_vision_override(cfg, provider, model)
     if override is not None:
         return override
     if not provider or not model:
         return None
-    caps = None
     try:
         from agent.models_dev import get_model_capabilities
         caps = get_model_capabilities(provider, model)
     except Exception as exc:  # pragma: no cover - defensive
         logger.debug("image_routing: caps lookup failed for %s:%s — %s", provider, model, exc)
-    if caps is not None:
-        return bool(caps.supports_vision)
-
-    base_url = _resolve_inference_base_url(cfg, provider)
-    if not base_url and (provider or "").strip().lower() == "ollama":
-        base_url = "http://localhost:11434/v1"
-    if _should_probe_ollama_vision(provider, base_url):
-        try:
-            from agent.model_metadata import query_ollama_supports_vision
-
-            ollama_vision = query_ollama_supports_vision(model, base_url)
-            if ollama_vision is not None:
-                return ollama_vision
-        except Exception as exc:  # pragma: no cover - defensive
-            logger.debug(
-                "image_routing: ollama vision probe failed for %s:%s — %s",
-                provider,
-                model,
-                exc,
-            )
-    return None
+        return None
+    if caps is None:
+        return None
+    return bool(caps.supports_vision)
 
 
 def decide_image_input_mode(
     provider: str,
     model: str,
     cfg: Optional[Dict[str, Any]],
-    *,
-    requested_provider: str = "",
 ) -> str:
     """Return ``"native"`` or ``"text"`` for the given turn.
 
@@ -471,7 +324,6 @@ def decide_image_input_mode(
       provider: active inference provider ID (e.g. ``"anthropic"``, ``"openrouter"``).
       model:    active model slug as it would be sent to the provider.
       cfg:      loaded config.yaml dict, or None. When None, behaves as auto.
-      requested_provider: provider identity before runtime canonicalization.
     """
     mode_cfg = "auto"
     if isinstance(cfg, dict):
@@ -484,25 +336,13 @@ def decide_image_input_mode(
     if mode_cfg == "text":
         return "text"
 
-    # auto: prefer native vision when the main model supports it. An
-    # explicit auxiliary.vision config acts as a *fallback* for text-only
-    # main models — it should not preempt native vision on a model that
-    # can natively inspect the pixels (issue #29135).
-    if requested_provider:
-        supports = _lookup_supports_vision(
-            provider,
-            model,
-            cfg,
-            requested_provider=requested_provider,
-        )
-    else:
-        # Keep the long-standing three-argument call contract for callers and
-        # tests that replace the capability lookup hook.
-        supports = _lookup_supports_vision(provider, model, cfg)
-    if supports is True:
-        return "native"
+    # auto
     if _explicit_aux_vision_override(cfg):
         return "text"
+
+    supports = _lookup_supports_vision(provider, model, cfg)
+    if supports is True:
+        return "native"
     return "text"
 
 
@@ -548,98 +388,14 @@ def _sniff_mime_from_bytes(raw: bytes) -> Optional[str]:
     # BMP: "BM"
     if raw.startswith(b"BM"):
         return "image/bmp"
-    # ISO-BMFF family (HEIC/HEIF/AVIF): bytes 4..8 == 'ftyp', major brand at 8..12
-    if len(raw) >= 12 and raw[4:8] == b"ftyp":
-        brand = raw[8:12]
-        if brand in {b"avif", b"avis"}:
-            return "image/avif"
-        if brand in {
-            b"heic", b"heix", b"hevc", b"hevx",
-            b"mif1", b"msf1", b"heim", b"heis",
-        }:
-            return "image/heic"
-    # TIFF: II*\0 (little-endian) or MM\0* (big-endian)
-    if raw[:4] in {b"II*\x00", b"MM\x00*"}:
-        return "image/tiff"
-    # ICO: 00 00 01 00 (reserved=0, type=1=icon)
-    if raw[:4] == b"\x00\x00\x01\x00":
-        return "image/x-icon"
-    # SVG: text-based, look for an <svg tag near the start (skip BOM/whitespace)
-    head = raw[:512].lstrip().lower()
-    if head.startswith(b"<?xml") or head.startswith(b"<svg"):
-        if b"<svg" in head:
-            return "image/svg+xml"
+    # HEIC/HEIF: ftypheic / ftypheix / ftypmif1 / ftypmsf1 etc.
+    if len(raw) >= 12 and raw[4:8] == b"ftyp" and raw[8:12] in {
+        b"heic", b"heix", b"hevc", b"hevx", b"mif1", b"msf1", b"heim", b"heis",
+    }:
+        return "image/heic"
     return None
 
 
-# Formats every major vision provider (Anthropic, OpenAI, Gemini, Bedrock)
-# accepts natively. Anything outside this set has to be transcoded to PNG
-# before we declare media_type, otherwise the provider returns HTTP 400
-# ("Could not process image" / "Unsupported image media type") and the
-# whole turn fails with no salvage path.
-#
-# Discord (and a few other chat platforms) freely accept attachments in
-# formats outside this set -- AVIF screenshots from Chromium, HEIC from
-# iPhones, TIFF from scanners, BMP from old Windows tools, ICO -- so users
-# do hit this in practice. SVG is vector and Pillow cannot rasterize it;
-# it is skipped (logged) rather than transcoded.
-_UNIVERSALLY_SUPPORTED_MIMES = frozenset({
-    "image/png", "image/jpeg", "image/gif", "image/webp",
-})
-
-
-def _transcode_to_png(raw: bytes) -> Optional[bytes]:
-    """Decode arbitrary image bytes with Pillow and re-encode as PNG.
-
-    Returns None if Pillow isn't installed or can't decode the input
-    (rare formats, corrupted bytes, missing optional decoder plugin for
-    HEIC/AVIF, or vector formats like SVG). Caller falls back to skipping
-    the image so the rest of the turn still works.
-
-    HEIC/HEIF and AVIF need optional Pillow plugins; we try to register
-    them on demand and swallow ImportError so a missing plugin just
-    looks like 'Pillow can't decode this' rather than crashing.
-    """
-    try:
-        from PIL import Image
-    except ImportError:
-        logger.info(
-            "image_routing: Pillow not installed; cannot transcode "
-            "non-standard image format to PNG. Install with `pip install Pillow` "
-            "(and `pillow-heif` / `pillow-avif-plugin` for those formats)."
-        )
-        return None
-    # Optional plugin registration. Silent on failure: an unsupported
-    # format will just fall through to Image.open raising below.
-    try:
-        import pillow_heif  # type: ignore
-
-        pillow_heif.register_heif_opener()
-    except Exception:
-        pass
-    try:
-        import pillow_avif  # type: ignore  # noqa: F401  -- registers AVIF on import
-    except Exception:
-        pass
-    try:
-        from io import BytesIO
-
-        with Image.open(BytesIO(raw)) as im:
-            # Pick an output mode PNG can serialise. Anything other than
-            # the standard set gets normalised to RGBA so transparency is
-            # preserved where the source had it.
-            if im.mode not in {"RGB", "RGBA", "L", "LA", "P"}:
-                im = im.convert("RGBA")
-            buf = BytesIO()
-            im.save(buf, format="PNG", optimize=False)
-            return buf.getvalue()
-    except Exception as exc:
-        logger.info(
-            "image_routing: Pillow could not transcode image to PNG -- %s", exc
-        )
-        return None
-
-
 def _guess_mime(path: Path, raw: Optional[bytes] = None) -> str:
     """Return image MIME type for *path*.
 
@@ -675,52 +431,15 @@ def _file_to_data_url(path: Path) -> Optional[str]:
     accept large images (OpenAI 49 MB+, Gemini 100 MB) don't pay a silent
     quality tax just because one other provider is stricter.
 
-    Format compatibility IS handled here: if the sniffed MIME isn't one
-    of ``_UNIVERSALLY_SUPPORTED_MIMES`` (i.e. it's something like AVIF,
-    HEIC, BMP, TIFF, or ICO that some providers reject outright), we
-    transcode to PNG with Pillow before declaring media_type. This fixes
-    the user-visible "Could not process image" HTTP 400 from Anthropic on
-    Discord-attached AVIF/HEIC/BMP files.
-
-    Returns None if the file can't be read OR if the format isn't
-    universally supported AND Pillow can't transcode it (Pillow missing,
-    HEIC/AVIF plugin missing, vector format like SVG, corrupt bytes). The
-    caller reports those paths in ``skipped`` and the rest of the turn
-    proceeds.
+    Returns None only if the file can't be read (missing, permission
+    denied, etc.); the caller reports those paths in ``skipped``.
     """
-    try:
-        from agent.file_safety import raise_if_read_blocked
-
-        raise_if_read_blocked(str(path))
-    except ValueError as exc:
-        logger.warning("image_routing: blocked local image attachment %s -- %s", path, exc)
-        return None
-    except Exception:
-        # Keep attachment routing best-effort if the guard itself is unavailable.
-        pass
-
     try:
         raw = path.read_bytes()
     except Exception as exc:
         logger.warning("image_routing: failed to read %s — %s", path, exc)
         return None
     mime = _guess_mime(path, raw=raw)
-    if mime not in _UNIVERSALLY_SUPPORTED_MIMES:
-        transcoded = _transcode_to_png(raw)
-        if transcoded is None:
-            logger.warning(
-                "image_routing: %s is %s which is not accepted by all major "
-                "vision providers and could not be transcoded to PNG; "
-                "skipping this attachment.",
-                path, mime,
-            )
-            return None
-        logger.info(
-            "image_routing: transcoded %s (%s) -> image/png for provider compatibility",
-            path.name, mime,
-        )
-        raw = transcoded
-        mime = "image/png"
     b64 = base64.b64encode(raw).decode("ascii")
     return f"data:{mime};base64,{b64}"
 

agent/insights.py

@@ -17,7 +17,6 @@ Usage:
 """
 
 import json
-import sqlite3
 import time
 from collections import Counter, defaultdict
 from datetime import datetime
@@ -142,8 +141,8 @@ class InsightsEngine:
             }
 
         # Compute insights
-        models = self._compute_model_breakdown(sessions, cutoff, source)
-        overview = self._compute_overview(sessions, message_stats, models)
+        overview = self._compute_overview(sessions, message_stats)
+        models = self._compute_model_breakdown(sessions)
         platforms = self._compute_platform_breakdown(sessions)
         tools = self._compute_tool_breakdown(tool_usage)
         skills = self._compute_skill_breakdown(skill_usage)
@@ -173,7 +172,7 @@ class InsightsEngine:
                      "message_count, tool_call_count, input_tokens, output_tokens, "
                      "cache_read_tokens, cache_write_tokens, billing_provider, "
                      "billing_base_url, billing_mode, estimated_cost_usd, "
-                     "actual_cost_usd, cost_status, cost_source, api_call_count")
+                     "actual_cost_usd, cost_status, cost_source")
 
     # Pre-computed query strings — f-string evaluated once at class definition,
     # not at runtime, so no user-controlled value can alter the query structure.
@@ -400,12 +399,7 @@ class InsightsEngine:
     # Computation
     # =========================================================================
 
-    def _compute_overview(
-        self,
-        sessions: List[Dict],
-        message_stats: Dict,
-        models: Optional[List[Dict]] = None,
-    ) -> Dict:
+    def _compute_overview(self, sessions: List[Dict], message_stats: Dict) -> Dict:
         """Compute high-level overview statistics."""
         total_input = sum(s.get("input_tokens") or 0 for s in sessions)
         total_output = sum(s.get("output_tokens") or 0 for s in sessions)
@@ -437,21 +431,6 @@ class InsightsEngine:
             else:
                 models_without_pricing.add(display)
 
-        if models:
-            total_cost = sum(float(m.get("cost") or 0.0) for m in models)
-            # Token totals likewise: the per-model breakdown includes
-            # auxiliary usage rows (vision/compression/titles — task
-            # dimension in session_model_usage, #23270) plus reconciled
-            # residuals, while the sessions counters carry main-loop usage
-            # only. Summing the breakdown keeps overview totals consistent
-            # with the per-model table and stops `hermes insights`
-            # undercounting aux spend (#58592, #9979).
-            total_input = sum(int(m.get("input_tokens") or 0) for m in models)
-            total_output = sum(int(m.get("output_tokens") or 0) for m in models)
-            total_cache_read = sum(int(m.get("cache_read_tokens") or 0) for m in models)
-            total_cache_write = sum(int(m.get("cache_write_tokens") or 0) for m in models)
-            total_tokens = total_input + total_output + total_cache_read + total_cache_write
-
         # Session duration stats (guard against negative durations from clock drift)
         durations = []
         for s in sessions:
@@ -494,189 +473,39 @@ class InsightsEngine:
             "included_cost_sessions": included_cost_sessions,
         }
 
-    _GET_MODEL_USAGE_WITH_SOURCE = (
-        "SELECT u.session_id, u.model, u.billing_provider, u.billing_base_url,"
-        " u.api_call_count, u.input_tokens, u.output_tokens,"
-        " u.cache_read_tokens, u.cache_write_tokens, u.reasoning_tokens,"
-        " u.estimated_cost_usd, u.actual_cost_usd, u.cost_status,"
-        " u.cost_source, u.billing_mode"
-        " FROM session_model_usage u"
-        " JOIN sessions s ON s.id = u.session_id"
-        " WHERE s.started_at >= ? AND s.source = ?"
-    )
-    _GET_MODEL_USAGE_ALL = (
-        "SELECT u.session_id, u.model, u.billing_provider, u.billing_base_url,"
-        " u.api_call_count, u.input_tokens, u.output_tokens,"
-        " u.cache_read_tokens, u.cache_write_tokens, u.reasoning_tokens,"
-        " u.estimated_cost_usd, u.actual_cost_usd, u.cost_status,"
-        " u.cost_source, u.billing_mode"
-        " FROM session_model_usage u"
-        " JOIN sessions s ON s.id = u.session_id"
-        " WHERE s.started_at >= ?"
-    )
-
-    def _get_model_usage(self, cutoff: float, source: str = None) -> List[Dict]:
-        """Fetch per-model usage rows within the window (issue #51607).
-
-        Returns an empty list when the table is missing (e.g. a DB opened by
-        older code that never created it) so the caller can fall back to the
-        per-session aggregate.
-        """
-        try:
-            if source:
-                cursor = self._conn.execute(
-                    self._GET_MODEL_USAGE_WITH_SOURCE, (cutoff, source)
-                )
-            else:
-                cursor = self._conn.execute(self._GET_MODEL_USAGE_ALL, (cutoff,))
-            return [dict(row) for row in cursor.fetchall()]
-        except sqlite3.OperationalError:
-            return []
-
-    def _compute_model_breakdown(
-        self, sessions: List[Dict], cutoff: float, source: str = None
-    ) -> List[Dict]:
-        """Break down token usage and cost by model.
-
-        Tokens and cost are attributed per model from session_model_usage, so a
-        session that switched models mid-flight (via ``/model``) splits across
-        every model it used instead of dumping everything on the initial model
-        (issue #51607). Sessions without per-model rows — e.g. data written
-        before this table existed and not yet backfilled — fall back to their
-        single recorded (model, billing_provider) aggregate so nothing is lost.
-
-        Tool calls aren't tied to a specific API invocation, so they stay
-        attributed to the session's recorded model.
-        """
+    def _compute_model_breakdown(self, sessions: List[Dict]) -> List[Dict]:
+        """Break down usage by model."""
         model_data = defaultdict(lambda: {
-            "sessions": set(), "input_tokens": 0, "output_tokens": 0,
+            "sessions": 0, "input_tokens": 0, "output_tokens": 0,
             "cache_read_tokens": 0, "cache_write_tokens": 0,
-            "reasoning_tokens": 0, "total_tokens": 0, "api_calls": 0,
-            "tool_calls": 0, "cost": 0.0, "actual_cost": 0.0,
+            "total_tokens": 0, "tool_calls": 0, "cost": 0.0,
         })
 
-        def _accumulate(model, provider, base_url, session_id, inp, out,
-                        cache_read, cache_write, reasoning, *,
-                        stored_cost=None, actual_cost=None, cost_status=None):
-            model = model or "unknown"
+        for s in sessions:
+            model = s.get("model") or "unknown"
             # Normalize: strip provider prefix for display
             display_model = model.split("/")[-1] if "/" in model else model
-            d: Dict[str, Any] = model_data[display_model]
-            d["sessions"].add(session_id)
+            d = model_data[display_model]
+            d["sessions"] += 1
+            inp = s.get("input_tokens") or 0
+            out = s.get("output_tokens") or 0
+            cache_read = s.get("cache_read_tokens") or 0
+            cache_write = s.get("cache_write_tokens") or 0
             d["input_tokens"] += inp
             d["output_tokens"] += out
             d["cache_read_tokens"] += cache_read
             d["cache_write_tokens"] += cache_write
-            d["reasoning_tokens"] += reasoning
             d["total_tokens"] += inp + out + cache_read + cache_write
-            if stored_cost is None:
-                estimate, status = _estimate_cost(
-                    model, inp, out,
-                    cache_read_tokens=cache_read, cache_write_tokens=cache_write,
-                    provider=provider or None, base_url=base_url,
-                )
-            else:
-                estimate = float(stored_cost or 0.0)
-                status = cost_status or "unknown"
+            d["tool_calls"] += s.get("tool_call_count") or 0
+            estimate, status = _estimate_cost(s)
             d["cost"] += estimate
-            d["actual_cost"] += float(actual_cost or 0.0)
+            d["has_pricing"] = has_known_pricing(model, s.get("billing_provider"), s.get("billing_base_url"))
             d["cost_status"] = status
-            if has_known_pricing(model, provider or None, base_url):
-                d["has_pricing"] = True
-            else:
-                d.setdefault("has_pricing", False)
-            return display_model
 
-        usage_rows = self._get_model_usage(cutoff, source)
-        usage_totals = defaultdict(lambda: {
-            "input_tokens": 0, "output_tokens": 0, "cache_read_tokens": 0,
-            "cache_write_tokens": 0, "reasoning_tokens": 0,
-            "api_call_count": 0, "estimated_cost_usd": 0.0,
-            "actual_cost_usd": 0.0,
-        })
-        for r in usage_rows:
-            totals: Dict[str, Any] = usage_totals[r["session_id"]]
-            for key in (
-                "input_tokens", "output_tokens", "cache_read_tokens",
-                "cache_write_tokens", "reasoning_tokens", "api_call_count",
-            ):
-                totals[key] += r[key] or 0
-            totals["estimated_cost_usd"] += r["estimated_cost_usd"] or 0.0
-            totals["actual_cost_usd"] += r["actual_cost_usd"] or 0.0
-            d = _accumulate(
-                r["model"], r["billing_provider"], r.get("billing_base_url"),
-                r["session_id"], r["input_tokens"] or 0, r["output_tokens"] or 0,
-                r["cache_read_tokens"] or 0, r["cache_write_tokens"] or 0,
-                r["reasoning_tokens"] or 0,
-                stored_cost=(
-                    r["estimated_cost_usd"]
-                    if r.get("cost_status") or r.get("cost_source")
-                    else None
-                ),
-                actual_cost=r["actual_cost_usd"],
-                cost_status=r.get("cost_status"),
-            )
-            model_data[d]["api_calls"] += r["api_call_count"] or 0
-
-        # Reconcile against the aggregate row. This covers legacy sessions,
-        # interrupted migrations, and absolute cumulative updates without
-        # double-counting already-attributed route deltas.
-        for s in sessions:
-            totals = usage_totals[s["id"]]
-            inp = max(0, (s.get("input_tokens") or 0) - totals["input_tokens"])
-            out = max(0, (s.get("output_tokens") or 0) - totals["output_tokens"])
-            cache_read = max(
-                0, (s.get("cache_read_tokens") or 0) - totals["cache_read_tokens"]
-            )
-            cache_write = max(
-                0, (s.get("cache_write_tokens") or 0) - totals["cache_write_tokens"]
-            )
-            residual_cost = max(
-                0.0, float(s.get("estimated_cost_usd") or 0.0)
-                - totals["estimated_cost_usd"],
-            )
-            residual_actual = max(
-                0.0, float(s.get("actual_cost_usd") or 0.0)
-                - totals["actual_cost_usd"],
-            )
-            residual_calls = max(
-                0, (s.get("api_call_count") or 0) - totals["api_call_count"]
-            )
-            if not (
-                inp or out or cache_read or cache_write or residual_cost
-                or residual_actual or residual_calls
-            ):
-                continue
-            d = _accumulate(
-                s.get("model"), s.get("billing_provider"),
-                s.get("billing_base_url"), s["id"],
-                inp, out, cache_read, cache_write, 0,
-                stored_cost=residual_cost,
-                actual_cost=residual_actual,
-                cost_status=s.get("cost_status"),
-            )
-            residual_bucket: Dict[str, Any] = model_data[d]
-            residual_bucket["api_calls"] += residual_calls
-
-        # Tool calls are attributed by the session's recorded model.
-        for s in sessions:
-            tool_calls = s.get("tool_call_count") or 0
-            if not tool_calls:
-                continue
-            model = s.get("model") or "unknown"
-            display_model = model.split("/")[-1] if "/" in model else model
-            model_data[display_model]["tool_calls"] += tool_calls
-
-        result = []
-        for model, data in model_data.items():
-            entry = {"model": model, **data}
-            entry["sessions"] = len(data["sessions"])
-            # Models that surfaced only via tool-call attribution (no token
-            # rows) won't have these set by _accumulate — default them so the
-            # output shape is uniform for downstream/JSON consumers.
-            entry.setdefault("has_pricing", False)
-            entry.setdefault("cost_status", "unknown")
-            result.append(entry)
+        result = [
+            {"model": model, **data}
+            for model, data in model_data.items()
+        ]
         # Sort by tokens first, fall back to session count when tokens are 0
         result.sort(key=lambda x: (x["total_tokens"], x["sessions"]), reverse=True)
         return result

agent/kanban_stop.py

@@ -1,108 +0,0 @@
-"""Turn-end guard for kanban workers.
-
-Kanban workers must end with ``kanban_complete`` or ``kanban_block``. Models
-(especially GLM / Qwen families) sometimes narrate the next step
-("Let me write the report now") and stop with ``finish_reason=stop`` and no
-tool calls. Hermes treats that as a clean exit → ``rc=0`` → dispatcher
-``protocol_violation``.
-
-This module is policy-only: when a kanban worker tries to finish without a
-terminal board tool, return a bounded synthetic nudge so the conversation
-loop continues instead of exiting.
-"""
-
-from __future__ import annotations
-
-import os
-from typing import Any, Iterable, Optional
-
-
-_TERMINAL_KANBAN_TOOLS = frozenset({"kanban_complete", "kanban_block"})
-
-_DEFAULT_MAX_ATTEMPTS = 2
-
-
-def kanban_stop_nudge_enabled() -> bool:
-    """Return whether the kanban stop-guard is active for this process.
-
-    On when ``HERMES_KANBAN_TASK`` is set (dispatcher-spawned worker), unless
-    ``HERMES_KANBAN_STOP_NUDGE`` explicitly disables it.
-    """
-    env = os.environ.get("HERMES_KANBAN_STOP_NUDGE")
-    if env is not None and env.strip().lower() in {"0", "false", "no", "off"}:
-        return False
-    task = (os.environ.get("HERMES_KANBAN_TASK") or "").strip()
-    return bool(task)
-
-
-def _tool_call_name(tc: Any) -> str:
-    if isinstance(tc, dict):
-        fn = tc.get("function")
-        if isinstance(fn, dict):
-            return str(fn.get("name") or "")
-        return str(tc.get("name") or "")
-    fn = getattr(tc, "function", None)
-    if fn is not None:
-        return str(getattr(fn, "name", "") or "")
-    return str(getattr(tc, "name", "") or "")
-
-
-def session_called_kanban_terminal(messages: Iterable[dict] | None) -> bool:
-    """True if this conversation already invoked a terminal kanban tool."""
-    if not messages:
-        return False
-    for msg in messages:
-        if not isinstance(msg, dict):
-            continue
-        role = msg.get("role")
-        if role == "assistant":
-            for tc in msg.get("tool_calls") or []:
-                if _tool_call_name(tc) in _TERMINAL_KANBAN_TOOLS:
-                    return True
-        elif role == "tool":
-            name = str(msg.get("name") or "")
-            if name in _TERMINAL_KANBAN_TOOLS:
-                return True
-    return False
-
-
-def build_kanban_stop_nudge(
-    *,
-    messages: Iterable[dict] | None = None,
-    attempts: int = 0,
-    max_attempts: int = _DEFAULT_MAX_ATTEMPTS,
-    task_id: Optional[str] = None,
-) -> Optional[str]:
-    """Return a synthetic follow-up when a kanban worker exits without a terminal tool.
-
-    Returns ``None`` when the guard should not fire (not a kanban worker,
-    already completed/blocked, or nudge budget exhausted).
-    """
-    if not kanban_stop_nudge_enabled():
-        return None
-    if attempts >= max_attempts:
-        return None
-    if session_called_kanban_terminal(messages):
-        return None
-
-    tid = (task_id or os.environ.get("HERMES_KANBAN_TASK") or "").strip() or "this task"
-    return (
-        "[System: You are a Hermes kanban worker. A plain-text reply is NOT a "
-        "terminal state for the board.\n\n"
-        f"Task `{tid}` is still `running`. Ending now without a board tool "
-        "causes a protocol violation (clean exit with no "
-        "`kanban_complete` / `kanban_block`).\n\n"
-        "Do this immediately in your next response — do not narrate intent:\n"
-        "1. Finish any remaining deliverable (write the required file(s) now).\n"
-        "2. Call `kanban_complete(summary=..., artifacts=[...])` if the work "
-        "is done, OR `kanban_block(reason=...)` if you are blocked.\n\n"
-        "Never end a turn with only a promise of future action. Repeated "
-        "protocol violations will block this task and require manual intervention.]"
-    )
-
-
-__all__ = [
-    "build_kanban_stop_nudge",
-    "kanban_stop_nudge_enabled",
-    "session_called_kanban_terminal",
-]

agent/learn_prompt.py

@@ -1,150 +0,0 @@
-#!/usr/bin/env python3
-"""``/learn`` — build the standards-guided prompt that turns whatever the user
-described into a reusable skill.
-
-``/learn`` is open-ended. The user can point it at anything they can describe:
-a directory of code, an API doc URL, a workflow they just walked the agent
-through in this conversation, or pasted notes. This module builds ONE prompt
-that instructs the live agent to:
-
-  1. Gather the sources the user named, using the tools it already has
-     (``read_file`` / ``search_files`` for dirs, ``web_extract`` for URLs, the
-     current conversation for "what I just did", the user's text for pasted
-     material).
-  2. Author a single ``SKILL.md`` via ``skill_manage`` that follows the Hermes
-     skill-authoring standards (description <=60 chars, the modern section
-     order, Hermes-tool framing, no invented commands).
-
-There is no separate distillation engine and no model-tool footprint: the
-agent does the work with its existing toolset, so this works identically on
-local, Docker, and remote terminal backends. Every surface (CLI ``/learn``,
-gateway ``/learn``, the dashboard "Learn a skill" panel) calls
-:func:`build_learn_prompt` and feeds the result to the agent as a normal turn.
-"""
-
-from __future__ import annotations
-
-# The house-style rules, distilled from AGENTS.md "Skill authoring standards
-# (HARDLINE)" and the hermes-agent-dev new-skill salvage reference. Embedded in
-# the prompt so the agent authors skills the way a maintainer would by hand.
-_AUTHORING_STANDARDS = """\
-Follow the Hermes skill-authoring standards exactly. These are the same
-HARDLINE rules a maintainer enforces in review:
-
-Frontmatter:
-- name: lowercase-hyphenated, <=64 chars, no spaces.
-- description: ONE sentence, **<=60 characters**, ends with a period. State the
-  capability, not the implementation. No marketing words (powerful,
-  comprehensive, seamless, advanced, robust). Do NOT repeat the skill name. If
-  the description contains a colon, wrap the whole value in double quotes.
-  This is the most-violated rule and it is NOT cosmetic: the system-prompt
-  skill index truncates the description to 60 chars and loads it every
-  session, so anything past char 60 is silently cut and never routes. After
-  you write the description, COUNT the characters; if it is over 60, cut it
-  down before saving — do not ship a sentence and hope.
-    Good (<=60): `Search arXiv papers by keyword, author, or ID.`
-    Bad (123):   `A comprehensive skill that lets the agent search arXiv for
-                  academic papers using keywords, authors, and categories.`
-- version: 0.1.0
-- author: always the literal value `Hermes`. NEVER fill it from the host
-  environment — the OS/login username (e.g. the `user=` line in your
-  environment hints), git config, or any identity you can probe must not be
-  written. Skills get shared and published, so an environment-derived name is
-  a privacy leak the user never opted into; the skill names itself as Hermes.
-- platforms: declare `[macos]`, `[linux]`, and/or `[windows]` IF the skill
-  uses OS-bound primitives (osascript/apt/systemctl => the matching OS; /proc,
-  os.setsid, signal.SIGKILL => linux; fcntl/termios => POSIX). Prefer fixing it
-  cross-platform first (tempfile.gettempdir(), pathlib.Path, psutil); gate only
-  when the dependency is genuinely platform-bound. Omit the field for portable
-  skills.
-- metadata.hermes.tags: a few Capitalized, Relevant, Tags.
-
-Body section order (omit a section only if it genuinely has no content):
-1. "# <Human Title>" then a 2-3 sentence intro: what it does, what it does NOT
-   do, and the key dependency stance (e.g. "stdlib only").
-2. "## When to Use" — bullet list of concrete trigger phrases.
-3. "## Prerequisites" — exact env vars, install steps, credentials.
-4. "## How to Run" — the canonical invocation, framed through Hermes tools.
-5. "## Quick Reference" — a flat command/endpoint list, no narration.
-6. "## Procedure" — numbered steps with copy-paste-exact commands.
-7. "## Pitfalls" — known limits, rate limits, things that look broken but aren't.
-8. "## Verification" — a single command/check that proves the skill worked.
-
-Hermes-tool framing (this is what makes it a skill, not shell docs):
-- Frame running scripts as "invoke through the `terminal` tool".
-- Reference Hermes tools by name in backticks: `terminal`, `read_file`,
-  `write_file`, `search_files`, `patch`, `web_extract`, `web_search`,
-  `vision_analyze`, `browser_navigate`, `delegate_task`, `image_generate`,
-  `text_to_speech`, `cronjob`, `memory`, `skill_view`, `execute_code`.
-- Do NOT name shell utilities the agent already has wrapped: say `read_file`
-  not cat/head/tail, `search_files` not grep/rg/find/ls, `patch` not sed/awk,
-  `web_extract` not curl-to-scrape, `write_file` not echo>file or heredocs.
-- Third-party CLIs (ffmpeg, gh, an SDK) are fine inside a script file, but the
-  prose still frames them as "invoke through the `terminal` tool". If the
-  skill needs an MCP server, name it and document its setup in Prerequisites.
-
-Quality bar:
-- Prefer exact commands, endpoint URLs, function signatures, and config keys
-  that appear VERBATIM in the source. NEVER invent flags, paths, or APIs — if
-  you didn't see it in the source, don't write it.
-- Keep it tight and scannable: ~100 lines for a simple skill, ~200 for a
-  complex one. Don't re-paste the source docs.
-- Don't write a router/index/hub skill that only points at other skills.
-- Larger scripts/parsers belong in a `scripts/` file (add via
-  `skill_manage` write_file), referenced from SKILL.md by relative path — not
-  inlined for the agent to re-type every run. References go in `references/`,
-  templates in `templates/`."""
-
-
-def build_learn_prompt(user_request: str) -> str:
-    """Build the agent prompt for an open-ended ``/learn`` request.
-
-    Args:
-        user_request: the free-text the user gave after ``/learn`` — a
-            description of the workflow, paths, URLs, or "what I just did".
-
-    Returns:
-        A complete instruction the agent runs as a normal turn. The agent
-        gathers the described sources with its existing tools and authors the
-        skill via ``skill_manage``.
-    """
-    req = (user_request or "").strip()
-    if not req:
-        req = (
-            "the workflow we just went through in this conversation — review "
-            "the steps taken and distill them into a reusable skill"
-        )
-
-    return (
-        "[/learn] The user wants you to learn a reusable skill from the "
-        "request below, and save it.\n\n"
-        f"THE REQUEST:\n{req}\n\n"
-        "The request is open-ended and may mix two kinds of content, in any "
-        "order: SOURCES to gather (directories, file paths, URLs, \"what we "
-        "just did\", pasted notes) AND REQUIREMENTS that shape the skill "
-        "(what to focus on, what to leave out, scope, naming, the angle to "
-        "take). Treat EVERY part of the request as load-bearing. In "
-        "particular, prose that comes after a path or link is NOT incidental "
-        "— it is the user telling you what they want from that source. A "
-        "request like `<url> focus on the auth flow, skip the deprecated "
-        "endpoints` means: gather the URL AND honor \"focus on auth, skip "
-        "deprecated\" as authoring requirements. Never fetch the first source "
-        "and ignore the rest.\n\n"
-        "Do this:\n"
-        "1. Gather every source the user named, using the tools you already "
-        "have — `read_file`/`search_files` for local files or directories, "
-        "`web_extract` for URLs, the current conversation history if they "
-        "referred to something you just did, and the text they pasted as-is. "
-        "If the request is ambiguous about scope, make a reasonable choice "
-        "and note it; do not stall.\n"
-        "1b. Apply every requirement, focus, and constraint in the request to "
-        "the skill you author — these govern what the SKILL.md covers and "
-        "emphasizes, not just which sources you read.\n"
-        "2. Author ONE SKILL.md and save it with the `skill_manage` tool "
-        "(action=\"create\"). Pick a sensible category. If the procedure needs "
-        "a non-trivial script, add it under the skill's `scripts/` with "
-        "`skill_manage` write_file and reference it by relative path.\n\n"
-        f"{_AUTHORING_STANDARDS}\n\n"
-        "When done, tell the user the skill name, its category, and a "
-        "one-line summary of what it captured."
-    )

agent/learning_graph.py

@@ -1,328 +0,0 @@
-"""Assemble the "learning made visible" graph for desktop.
-
-This graph is intentionally scoped to what a user actually learns over time:
-- non-base, learned/profile skills (agent-created or used),
-- memory chunks from ``MEMORY.md`` / ``USER.md`` as first-class nodes.
-
-Skill links come from declared ``related_skills``. Memory-to-skill links are
-derived from lexical overlap so the graph can answer "which learned skills are
-connected to the things I remember?".
-
-Run as a module to print edge-density stats against real data:
-
-    python -m agent.learning_graph
-"""
-
-from __future__ import annotations
-
-import json
-import re
-from dataclasses import dataclass, field
-from datetime import datetime, timezone
-from pathlib import Path
-from typing import Any, Optional
-
-from hermes_constants import get_hermes_home
-
-
-@dataclass
-class SkillNode:
-    name: str
-    category: str
-    source: str = "profile"
-    timestamp: Optional[int] = None
-    use_count: int = 0
-    state: str = "active"
-    created_by: Optional[str] = None
-    pinned: bool = False
-    related: list[str] = field(default_factory=list)
-
-
-def _frontmatter(text: str) -> dict[str, Any]:
-    try:
-        from agent.skill_utils import parse_frontmatter
-
-        fm, _ = parse_frontmatter(text)
-        return fm or {}
-    except Exception:
-        return {}
-
-
-def _hermes_meta(fm: dict[str, Any]) -> dict[str, Any]:
-    """``metadata.hermes`` as a dict, tolerant of the string-valued frontmatter
-    that ``parse_frontmatter``'s malformed-YAML fallback produces."""
-    meta = fm.get("metadata")
-    hermes = meta.get("hermes") if isinstance(meta, dict) else None
-    return hermes if isinstance(hermes, dict) else {}
-
-
-def _related(fm: dict[str, Any]) -> list[str]:
-    raw = fm.get("related_skills") or _hermes_meta(fm).get("related_skills")
-    if isinstance(raw, list):
-        return [str(r).strip() for r in raw if str(r).strip()]
-    if isinstance(raw, str):
-        return [r.strip() for r in raw.strip("[]").split(",") if r.strip()]
-    return []
-
-
-def _category(fm: dict[str, Any], skill_md: Path) -> str:
-    cat = fm.get("category") or _hermes_meta(fm).get("category")
-    if cat:
-        return str(cat)
-    # …/skills/<category>/<skill>/SKILL.md
-    parts = skill_md.parts
-    return parts[-3] if len(parts) >= 3 else "general"
-
-
-def _iter_skill_files(roots: list[tuple[str, Path]]):
-    for source, root in roots:
-        if root.exists():
-            for path in root.rglob("SKILL.md"):
-                yield source, path
-
-
-def _load_usage() -> dict[str, dict[str, Any]]:
-    try:
-        from tools.skill_usage import load_usage
-
-        return load_usage()
-    except Exception:
-        path = get_hermes_home() / "skills" / ".usage.json"
-        try:
-            return json.loads(path.read_text(encoding="utf-8"))
-        except Exception:
-            return {}
-
-
-def _to_int_ts(value: Any) -> Optional[int]:
-    try:
-        if value is None:
-            return None
-        if isinstance(value, (int, float)):
-            return int(value)
-        s = str(value).strip()
-        if not s:
-            return None
-        try:
-            return int(float(s))
-        except ValueError:
-            parsed = datetime.fromisoformat(s.replace("Z", "+00:00"))
-            if parsed.tzinfo is None:
-                parsed = parsed.replace(tzinfo=timezone.utc)
-            return int(parsed.timestamp())
-    except Exception:
-        return None
-
-
-def _usage_timestamp(rec: dict[str, Any]) -> Optional[int]:
-    for key in ("last_activity_at", "last_used_at", "last_viewed_at", "last_patched_at", "created_at"):
-        ts = _to_int_ts(rec.get(key))
-        if ts is not None:
-            return ts
-    return None
-
-
-def build_skill_nodes(skill_roots: list[tuple[str, Path]]) -> dict[str, SkillNode]:
-    usage = _load_usage()
-    nodes: dict[str, SkillNode] = {}
-
-    for source, skill_md in _iter_skill_files(skill_roots):
-        if any(p in {".archive", ".hub", "node_modules", ".git"} for p in skill_md.parts):
-            continue
-        try:
-            fm = _frontmatter(skill_md.read_text(encoding="utf-8")[:4000])
-        except OSError:
-            continue
-        name = str(fm.get("name") or skill_md.parent.name).strip()
-        if not name or name in nodes:
-            continue
-        rec = usage.get(name, {})
-        last_activity = _usage_timestamp(rec)
-        file_ts = _to_int_ts(skill_md.stat().st_mtime)
-        nodes[name] = SkillNode(
-            name=name,
-            category=_category(fm, skill_md),
-            source=source,
-            timestamp=last_activity or file_ts,
-            use_count=int(rec.get("use_count", 0) or 0),
-            state=str(rec.get("state", "active") or "active"),
-            created_by=rec.get("created_by"),
-            pinned=bool(rec.get("pinned", False)),
-            related=_related(fm),
-        )
-    return nodes
-
-
-def build_edges(nodes: dict[str, SkillNode]) -> list[tuple[str, str]]:
-    """Undirected related_skills edges where BOTH endpoints exist (deduped)."""
-    seen: set[tuple[str, str]] = set()
-    edges: list[tuple[str, str]] = []
-    for node in nodes.values():
-        for target in node.related:
-            if target in nodes and target != node.name:
-                a, b = sorted((node.name, target))
-                key = (a, b)
-                if key not in seen:
-                    seen.add(key)
-                    edges.append(key)
-    return edges
-
-
-def density_stats(nodes: dict[str, SkillNode], edges: list[tuple[str, str]]) -> dict[str, Any]:
-    linked: set[str] = set()
-    for a, b in edges:
-        linked.add(a)
-        linked.add(b)
-    cats: dict[str, int] = {}
-    for n in nodes.values():
-        cats[n.category] = cats.get(n.category, 0) + 1
-    n = len(nodes) or 1
-    return {
-        "nodes": len(nodes),
-        "related_edges": len(edges),
-        "edges_per_node": round(len(edges) / n, 3),
-        "linked_nodes": len(linked),
-        "isolated_pct": round(100 * (n - len(linked)) / n, 1),
-        "categories": len(cats),
-        "agent_created": sum(1 for x in nodes.values() if x.created_by == "agent"),
-        "used": sum(1 for x in nodes.values() if x.use_count > 0),
-        "top_categories": sorted(cats.items(), key=lambda kv: -kv[1])[:8],
-    }
-
-
-def _memory_cards() -> list[dict[str, Any]]:
-    """Freeform memory as readable cards.
-
-    ``MEMORY.md`` / ``USER.md`` are prose split on bare ``§`` separators; each
-    chunk becomes one card. Every chunk is surfaced — the graph shows everything.
-    """
-    base = get_hermes_home() / "memories"
-    cards: list[dict[str, Any]] = []
-    for fname, source in (("MEMORY.md", "memory"), ("USER.md", "profile")):
-        path = base / fname
-        try:
-            text = path.read_text(encoding="utf-8").strip()
-            file_ts = _to_int_ts(path.stat().st_mtime)
-        except OSError:
-            continue
-        for chunk_idx, chunk in enumerate(c.strip() for c in text.split("\n§\n")):
-            if not chunk:
-                continue
-            first = chunk.splitlines()[0].strip().lstrip("# ").strip()
-            cards.append(
-                {
-                    "source": source,
-                    "timestamp": file_ts + chunk_idx if file_ts is not None else None,
-                    "title": (first[:80] + "…") if len(first) > 80 else first,
-                    "body": chunk[:1200],
-                }
-            )
-    return cards
-
-
-def _tokenize(text: str) -> set[str]:
-    return {t for t in re.split(r"[^a-z0-9]+", text.lower()) if len(t) >= 3}
-
-
-def _memory_skill_edges(memory_cards: list[dict[str, Any]], skills: list[SkillNode]) -> list[tuple[str, str]]:
-    edges: list[tuple[str, str]] = []
-    skill_meta = [(s, _tokenize(s.name), s.name.lower()) for s in skills]
-    for idx, card in enumerate(memory_cards):
-        mem_id = f"memory:{card['source']}:{idx}"
-        text = f"{card.get('title', '')}\n{card.get('body', '')}".lower()
-        text_tokens = _tokenize(text)
-        scored: list[tuple[int, str]] = []
-        for skill, tokens, skill_name_lower in skill_meta:
-            score = 0
-            if skill_name_lower in text:
-                score += 6
-            score += len(tokens & text_tokens)
-            if score > 0:
-                scored.append((score, skill.name))
-        scored.sort(key=lambda x: (-x[0], x[1]))
-        for _, skill_name in scored[:4]:
-            edges.append((mem_id, skill_name))
-    return edges
-
-
-def _skill_roots() -> list[tuple[str, Path]]:
-    repo = Path(__file__).resolve().parent.parent
-    home_skills = get_hermes_home() / "skills"
-    return [("base", repo / "skills"), ("profile", home_skills)]
-
-
-def build_learning_graph() -> dict[str, Any]:
-    """Full payload for the desktop learning panel.
-
-    Focus on what is profile-learned and actionable:
-    - skills that are NOT base-installed and show real learning signal
-      (agent-created or used),
-    - memory chunks as first-class graph nodes connected to those learned skills.
-    """
-    all_skills = build_skill_nodes(_skill_roots())
-    learned_skills = {
-        name: node
-        for name, node in all_skills.items()
-        if node.source != "base" and (node.created_by == "agent" or node.use_count > 0)
-    }
-    skill_edges = build_edges(learned_skills)
-    memory_cards = _memory_cards()
-    memory_edges = _memory_skill_edges(memory_cards, list(learned_skills.values()))
-
-    edges = skill_edges + memory_edges
-    clusters: dict[str, int] = {}
-    for node in learned_skills.values():
-        clusters[node.category] = clusters.get(node.category, 0) + 1
-    if memory_cards:
-        clusters["memory"] = len(memory_cards)
-
-    graph_nodes = [
-        {
-            "id": n.name,
-            "label": n.name,
-            "kind": "skill",
-            "timestamp": n.timestamp,
-            "category": n.category,
-            "useCount": n.use_count,
-            "state": n.state,
-            "createdBy": n.created_by,
-            "pinned": n.pinned,
-        }
-        for n in learned_skills.values()
-    ]
-    for i, card in enumerate(memory_cards):
-        graph_nodes.append(
-            {
-                "id": f"memory:{card['source']}:{i}",
-                "label": card["title"],
-                "kind": "memory",
-                "memorySource": card["source"],
-                "timestamp": card.get("timestamp"),
-                "category": "memory",
-                "useCount": 0,
-                "state": "active",
-                "createdBy": "memory",
-                "pinned": False,
-            }
-        )
-
-    return {
-        "nodes": graph_nodes,
-        "edges": [{"source": a, "target": b} for a, b in edges],
-        "clusters": [
-            {"category": c, "count": n}
-            for c, n in sorted(clusters.items(), key=lambda kv: -kv[1])
-        ],
-        "memory": memory_cards,
-        "stats": {
-            **density_stats(learned_skills, skill_edges),
-            "memory_nodes": len(memory_cards),
-            "memory_skill_edges": len(memory_edges),
-            "learned_skills": len(learned_skills),
-        },
-    }
-
-
-if __name__ == "__main__":
-    nodes = build_skill_nodes(_skill_roots())
-    print(json.dumps(density_stats(nodes, build_edges(nodes)), indent=2))

agent/learning_graph_render.py

@@ -1,659 +0,0 @@
-"""Terminal renderer for the learning timeline (learned skills + memories).
-
-The desktop app (``apps/desktop/src/app/starmap``) paints a GPU radial
-constellation; a terminal can't, so this is a *rendition* of the same data as a
-timeline bar chart — date rows, proportional skill/memory bars colored by the
-day's dominant category, and a cumulative trajectory sparkline — plus per-slice
-bucket metadata the TUI walks as a tree. The age gradient and complementary
-memory ink are ported from the desktop source, not guessed.
-
-Grids are emitted as style runs — ``[text, style, alpha, hex?]`` — so each
-consumer maps the semantic style + brightness onto its own palette; the
-optional 4th element overrides the base color (category heatmap). Pure,
-stdlib-only.
-"""
-
-from __future__ import annotations
-
-import math
-from datetime import datetime, timezone
-from typing import Any, Iterable, Optional
-
-# time-axis.ts LEAD_IN: the oldest node sits just off recency 0.
-LEAD_IN = 0.06
-
-# constants.ts AGE_GRADIENT — old quiet, recent bright.
-AGE_OLD_INK = 0.42
-AGE_MID_INK = 0.74
-AGE_NEW_INK = 0.95
-AGE_MID = 0.52
-
-# Style keys consumers map to base colors (brightness = the run alpha).
-STYLE_BG = "bg"
-STYLE_SKILL = "skill"
-STYLE_MEMORY = "memory"
-STYLE_LABEL = "label"
-STYLE_DIM = "dim"
-
-# Legend glyphs mirror NODE_SHAPE (skill = circle, memory = diamond).
-SKILL_GLYPH = "●"
-MEMORY_GLYPH = "◆"
-_LABEL_KEYS = tuple("123456789abc")
-
-Run = list  # [text, style, alpha, hex?]
-Row = list  # list[Run]
-Grid = list  # list[Row]
-
-
-def _to_ts(value: Any) -> Optional[float]:
-    try:
-        return None if value is None else float(value)
-    except (TypeError, ValueError):
-        return None
-
-
-def _clamp(v: float, lo: float, hi: float) -> float:
-    return lo if v < lo else hi if v > hi else v
-
-
-def _smoothstep(p: float) -> float:
-    p = _clamp(p, 0.0, 1.0)
-    return p * p * (3 - 2 * p)
-
-
-def recency_ink(rec: float) -> float:
-    """Port of geometry.ts ``recencyInk`` — smoothstep age → ink alpha."""
-    t = _clamp(rec, 0.0, 1.0)
-    if t <= AGE_MID:
-        return AGE_OLD_INK + (AGE_MID_INK - AGE_OLD_INK) * _smoothstep(t / AGE_MID)
-    return AGE_MID_INK + (AGE_NEW_INK - AGE_MID_INK) * _smoothstep((t - AGE_MID) / (1 - AGE_MID))
-
-
-def format_date(ts: Optional[float]) -> str:
-    if not ts:
-        return "unknown"
-    try:
-        dt = datetime.fromtimestamp(float(ts), tz=timezone.utc)
-        return f"{dt.day} {dt.strftime('%b %Y')}"
-    except (ValueError, OSError, OverflowError):
-        return "unknown"
-
-
-def compute_recency(nodes: list[dict[str, Any]]) -> dict[str, Any]:
-    """Port of time-axis.ts ``computeRecency`` (id → recency ratio, timed flag)."""
-    known = [t for t in (_to_ts(n.get("timestamp")) for n in nodes) if t is not None]
-    min_ts = min(known) if known else None
-    max_ts = max(known) if known else None
-    timed = min_ts is not None and max_ts is not None and max_ts > min_ts
-
-    ordered = sorted(
-        nodes,
-        key=lambda n: (
-            _to_ts(n.get("timestamp")) if _to_ts(n.get("timestamp")) is not None else math.inf,
-            str(n.get("id", "")),
-        ),
-    )
-    last = max(len(ordered) - 1, 1)
-    ord_ratio = {str(n.get("id", "")): (i / last if len(ordered) > 1 else 0.0) for i, n in enumerate(ordered)}
-
-    rec: dict[str, float] = {}
-    for n in nodes:
-        nid = str(n.get("id", ""))
-        ts = _to_ts(n.get("timestamp"))
-        if timed and ts is not None and min_ts is not None and max_ts is not None:
-            ratio = (ts - min_ts) / (max_ts - min_ts)
-        else:
-            ratio = ord_ratio.get(nid, 0.0)
-        rec[nid] = LEAD_IN + (1 - LEAD_IN) * _clamp(ratio, 0.0, 1.0)
-
-    return {"rec": rec, "timed": timed, "minTs": min_ts, "maxTs": max_ts}
-
-
-def _date_at(rec: dict[str, Any], reveal: float) -> Optional[float]:
-    if not rec.get("timed"):
-        return None
-    lo, hi = rec.get("minTs"), rec.get("maxTs")
-    if lo is None or hi is None:
-        return None
-    return round(lo + _clamp(reveal, 0, 1) * (hi - lo))
-
-
-# ── Color: ported from color.ts so memory ink + age fade match the desktop ──
-
-
-def hex_to_rgb(s: str) -> tuple[int, int, int]:
-    s = s.strip().lstrip("#")
-    if len(s) == 3:
-        s = "".join(c * 2 for c in s)
-    try:
-        return int(s[0:2], 16), int(s[2:4], 16), int(s[4:6], 16)
-    except (ValueError, IndexError):
-        return 255, 215, 0
-
-
-def rgb_to_hex(c: tuple) -> str:
-    return "#{:02X}{:02X}{:02X}".format(*(int(_clamp(v, 0, 255)) for v in c))
-
-
-def mix_rgb(a: tuple, b: tuple, t: float) -> tuple[int, int, int]:
-    p = _clamp(t, 0.0, 1.0)
-    return tuple(round(a[i] + (b[i] - a[i]) * p) for i in range(3))  # type: ignore[return-value]
-
-
-def _rgb_to_hsl(c: tuple) -> tuple[float, float, float]:
-    r, g, b = (x / 255 for x in c)
-    mx, mn = max(r, g, b), min(r, g, b)
-    light = (mx + mn) / 2
-    d = mx - mn
-    if not d:
-        return 0.0, 0.0, light
-    s = d / (2 - mx - mn) if light > 0.5 else d / (mx + mn)
-    if mx == r:
-        h = (g - b) / d + (6 if g < b else 0)
-    elif mx == g:
-        h = (b - r) / d + 2
-    else:
-        h = (r - g) / d + 4
-    return h * 60, s, light
-
-
-def _hsl_to_rgb(h: float, s: float, light: float) -> tuple[int, int, int]:
-    hue = ((h % 360) + 360) % 360
-    c = (1 - abs(2 * light - 1)) * s
-    x = c * (1 - abs(((hue / 60) % 2) - 1))
-    m = light - c / 2
-    if hue < 60:
-        r, g, b = c, x, 0.0
-    elif hue < 120:
-        r, g, b = x, c, 0.0
-    elif hue < 180:
-        r, g, b = 0.0, c, x
-    elif hue < 240:
-        r, g, b = 0.0, x, c
-    elif hue < 300:
-        r, g, b = x, 0.0, c
-    else:
-        r, g, b = c, 0.0, x
-    return round((r + m) * 255), round((g + m) * 255), round((b + m) * 255)
-
-
-def _complementary_ink(c: tuple) -> tuple[int, int, int]:
-    h, s, light = _rgb_to_hsl(c)
-    return _hsl_to_rgb(h + 165, max(s, 0.5), _clamp(light, 0.5, 0.7))
-
-
-def derive_palette(primary_hex: str, *, dark: bool = True) -> dict[str, str]:
-    """Port of color.ts ``computePalette`` (the bits a terminal needs)."""
-    primary = hex_to_rgb(primary_hex)
-    base = (255, 255, 255) if dark else (0, 0, 0)
-    bg = (8, 8, 12) if dark else (250, 250, 250)
-    return {
-        "primary": primary_hex,
-        # Memories are drillable → primary "clickable" ink; skills are dead-ends
-        # → muted complement.
-        "memory": rgb_to_hex(mix_rgb(primary, base, 0.12 if dark else 0.18)),
-        "skill": rgb_to_hex(mix_rgb(_complementary_ink(primary), bg, 0.45)),
-        "label": rgb_to_hex(mix_rgb(base, bg, 0.35)),
-        "dim": rgb_to_hex(mix_rgb(base, bg, 0.7)),
-        "bg": rgb_to_hex(bg),
-    }
-
-
-def _node_score(node: dict[str, Any], rec: float) -> float:
-    """Pick which visible objects deserve map markers + label rows."""
-    if node.get("kind") == "memory":
-        return 3.5 + rec
-    use = float(node.get("useCount", 0) or 0)
-    return rec * 2 + math.sqrt(max(0.0, use)) + (2.0 if node.get("pinned") else 0.0)
-
-
-def _node_label(node: dict[str, Any]) -> str:
-    text = str(node.get("label") or node.get("id") or "unknown").strip()
-    return text if len(text) <= 26 else text[:23].rstrip() + "…"
-
-
-def _node_meta(node: dict[str, Any]) -> str:
-    if node.get("kind") == "memory":
-        source = "profile memory" if node.get("memorySource") == "profile" else "memory"
-        return f"{source} · {format_date(_to_ts(node.get('timestamp')))}"
-    bits = [str(node.get("category") or "skill"), format_date(_to_ts(node.get("timestamp")))]
-    count = int(node.get("useCount", 0) or 0)
-    if count:
-        bits.append(f"x{count}")
-    if node.get("pinned"):
-        bits.append("pinned")
-    return " · ".join(bits)
-
-
-# ── Timeline chart frame ─────────────────────────────────────────────────────
-
-
-class _ChartBucket:
-    __slots__ = ("label", "ts", "skills", "memories", "nodes", "rec")
-
-    def __init__(self, label: str, ts: float):
-        self.label = label
-        self.ts = ts
-        self.skills = 0
-        self.memories = 0
-        self.nodes: list[dict[str, Any]] = []
-        self.rec = 1.0
-
-    @property
-    def total(self) -> int:
-        return self.skills + self.memories
-
-
-def _period_key(ts: float, granularity: str) -> tuple[int, ...]:
-    dt = datetime.fromtimestamp(ts, tz=timezone.utc)
-    if granularity == "day":
-        return (dt.year, dt.month, dt.day)
-    if granularity == "month":
-        return (dt.year, dt.month)
-    return (dt.year,)
-
-
-def _period_label(ts: float, granularity: str) -> str:
-    dt = datetime.fromtimestamp(ts, tz=timezone.utc)
-    if granularity == "day":
-        return f"{dt.day} {dt.strftime('%b')}"
-    if granularity == "month":
-        return dt.strftime("%b %Y")
-    return dt.strftime("%Y")
-
-
-def _build_chart_buckets(nodes: list[dict[str, Any]], rec: dict[str, Any], max_rows: int) -> list[_ChartBucket]:
-    """Timeline rows: finest date granularity that fits, oldest → newest."""
-    if not nodes:
-        return []
-    if not rec["timed"]:
-        ordered = sorted(nodes, key=lambda n: rec["rec"].get(str(n.get("id", "")), 0.0))
-        n_bins = min(max_rows, max(1, len(ordered)))
-        buckets = [_ChartBucket(f"#{i + 1}", float(i)) for i in range(n_bins)]
-        for node in ordered:
-            idx = int(_clamp(math.floor(rec["rec"].get(str(node.get("id", "")), 0.0) * n_bins), 0, n_bins - 1))
-            b = buckets[idx]
-            b.nodes.append(node)
-            if node.get("kind") == "memory":
-                b.memories += 1
-            else:
-                b.skills += 1
-        return buckets
-
-    chosen: Optional[list[_ChartBucket]] = None
-    for granularity in ("day", "month", "year"):
-        groups: dict[tuple[int, ...], _ChartBucket] = {}
-        for node in nodes:
-            ts = _to_ts(node.get("timestamp"))
-            if ts is None:
-                continue
-            key = _period_key(ts, granularity)
-            bucket = groups.get(key)
-            if bucket is None:
-                bucket = _ChartBucket(_period_label(ts, granularity), ts)
-                groups[key] = bucket
-            bucket.nodes.append(node)
-            if node.get("kind") == "memory":
-                bucket.memories += 1
-            else:
-                bucket.skills += 1
-        # For short spans, keep the useful day-by-day graph even when the caller
-        # asked for fewer rows; terminal scrollback is better than collapsing a
-        # month of activity into one unreadable bar.
-        if len(groups) <= max_rows or (granularity == "day" and len(groups) <= 32):
-            chosen = [groups[key] for key in sorted(groups)]
-            break
-
-    if chosen is None:
-        # If even yearly buckets overflow, fall back to even time bins.
-        min_ts, max_ts = rec.get("minTs"), rec.get("maxTs")
-        n_bins = max(1, max_rows)
-        chosen = []
-        for i in range(n_bins):
-            ts = min_ts + (i / max(1, n_bins - 1)) * (max_ts - min_ts) if min_ts and max_ts else float(i)
-            chosen.append(_ChartBucket(format_date(ts), ts))
-        for node in nodes:
-            r = rec["rec"].get(str(node.get("id", "")), 0.0)
-            idx = int(_clamp(math.floor(r * n_bins), 0, n_bins - 1))
-            b = chosen[idx]
-            b.nodes.append(node)
-            if node.get("kind") == "memory":
-                b.memories += 1
-            else:
-                b.skills += 1
-
-    min_ts, max_ts = rec.get("minTs"), rec.get("maxTs")
-    span = (max_ts - min_ts) if min_ts is not None and max_ts is not None and max_ts > min_ts else 0
-    for bucket in chosen:
-        bucket.rec = LEAD_IN + (1 - LEAD_IN) * ((bucket.ts - min_ts) / span) if span else 1.0
-    return chosen
-
-
-def _bucket_label_node(bucket: _ChartBucket) -> Optional[dict[str, Any]]:
-    if not bucket.nodes:
-        return None
-    return max(bucket.nodes, key=lambda node: _node_score(node, _to_ts(node.get("timestamp")) or bucket.ts))
-
-
-def _bucket_nodes(bucket: _ChartBucket, memory_lookup: Optional[dict[str, dict[str, Any]]] = None) -> list[dict[str, Any]]:
-    out: list[dict[str, Any]] = []
-    # Chronological within the slice so the TUI tree reads oldest → newest.
-    ordered = sorted(bucket.nodes, key=lambda n: _to_ts(n.get("timestamp")) or bucket.ts)
-    for node in ordered:
-        style = STYLE_MEMORY if node.get("kind") == "memory" else STYLE_SKILL
-        raw_label = str(node.get("label") or node.get("id") or "unknown").strip()
-        memory = (memory_lookup or {}).get(str(node.get("id", "")))
-        out.append(
-            {
-                "id": str(node.get("id", "")),
-                "glyph": MEMORY_GLYPH if node.get("kind") == "memory" else SKILL_GLYPH,
-                "label": _node_label(node),
-                "fullLabel": raw_label,
-                "meta": _node_meta(node),
-                "body": str(memory.get("body", "")) if memory else "",
-                "style": style,
-            }
-        )
-    return out
-
-
-def _bucket_rows(buckets: list[_ChartBucket], payload: dict[str, Any]) -> list[dict[str, Any]]:
-    cmap = category_color_map(payload)
-    memory_lookup = {
-        f"memory:{card.get('source')}:{idx}": card
-        for idx, card in enumerate(payload.get("memory", []) or [])
-        if isinstance(card, dict)
-    }
-    rows: list[dict[str, Any]] = []
-    for idx, bucket in enumerate(buckets):
-        cat = _bucket_category(bucket)
-        rows.append(
-            {
-                "index": idx,
-                "label": bucket.label,
-                "date": format_date(bucket.ts),
-                "skills": bucket.skills,
-                "memories": bucket.memories,
-                "total": bucket.total,
-                "category": cat,
-                "color": cmap.get(cat) if cat else None,
-                "nodes": _bucket_nodes(bucket, memory_lookup),
-            }
-        )
-    return rows
-
-
-def _category_counts(payload: dict[str, Any]) -> list[tuple[str, int]]:
-    clusters = [
-        (str(c.get("category")), int(c.get("count", 0)))
-        for c in payload.get("clusters", []) or []
-        if c.get("category") and c.get("category") != "memory"
-    ]
-    if clusters:
-        return clusters
-    counts: dict[str, int] = {}
-    for node in payload.get("nodes", []):
-        if node.get("kind") == "memory":
-            continue
-        cat = str(node.get("category") or "skill")
-        counts[cat] = counts.get(cat, 0) + 1
-    return sorted(counts.items(), key=lambda kv: (-kv[1], kv[0]))
-
-
-def category_color_map(payload: dict[str, Any]) -> dict[str, str]:
-    """Deterministic, evenly-spread hue per skill category (theme-independent)."""
-    clusters = _category_counts(payload)
-    n = max(1, len(clusters))
-    # Golden-angle hue spacing so adjacent categories never collide in color.
-    return {cat: rgb_to_hex(_hsl_to_rgb((i * 137.508) % 360, 0.55, 0.62)) for i, (cat, _c) in enumerate(clusters)}
-
-
-def category_legend(payload: dict[str, Any], limit: int = 4) -> list[dict[str, Any]]:
-    cmap = category_color_map(payload)
-    cats = _category_counts(payload)
-    shown = cats[:limit]
-    hidden = max(0, len(cats) - len(shown))
-    return [
-        {"glyph": "●", "color": cmap.get(cat, ""), "label": f"{cat} ({count})"}
-        for cat, count in shown
-    ] + ([{"glyph": "·", "color": "", "label": f"+{hidden}"}] if hidden else [])
-
-
-def _bucket_category(bucket: _ChartBucket) -> Optional[str]:
-    counts: dict[str, int] = {}
-    for node in bucket.nodes:
-        if node.get("kind") == "memory":
-            continue
-        cat = str(node.get("category") or "skill")
-        counts[cat] = counts.get(cat, 0) + 1
-    return max(counts, key=lambda k: counts[k]) if counts else None
-
-
-def _trajectory_row(buckets: list[_ChartBucket], width: int, reveal: float) -> Row:
-    """Cumulative learning curve as a compact star-path sparkline."""
-    if not buckets:
-        return []
-    total = sum(b.total for b in buckets) or 1
-    visible = int(_clamp(math.ceil(reveal * len(buckets)), 0, len(buckets)))
-    acc = 0
-    points: list[int] = []
-    for b in buckets[:visible]:
-        acc += b.total
-        points.append(round((acc / total) * (width - 1)))
-    cells = [" "] * width
-    last = 0
-    for p in points:
-        for x in range(min(last, p), max(last, p) + 1):
-            if 0 <= x < width and cells[x] == " ":
-                cells[x] = "·"
-        if 0 <= p < width:
-            cells[p] = "✦"
-        last = p
-    return [["trajectory ", STYLE_LABEL, 0.55], ["".join(cells), STYLE_SKILL, 0.48]]
-
-
-def render_graph(payload: dict[str, Any], *, cols: int = 80, rows: int = 16, reveal: float = 1.0) -> dict[str, Any]:
-    """Render one timeline frame at ``reveal`` (0→1).
-
-    Date rows with proportional skill/memory bars colored by the day's dominant
-    category, numbered markers tied to label rows, and a cumulative trajectory
-    sparkline underneath.
-    """
-    reveal = _clamp(reveal, 0.0, 1.0)
-    cols = max(44, cols)
-    rows = max(14, rows)
-    nodes = list(payload.get("nodes", []))
-    if not nodes:
-        placeholder = [["no learning yet — keep using Hermes and it maps out here", STYLE_DIM, 0.7]]
-        return {"grid": [placeholder], "date": "", "reveal": reveal, "visible": 0}
-
-    rec = compute_recency(nodes)
-    cmap = category_color_map(payload)
-    buckets = _build_chart_buckets(nodes, rec, max_rows=max(4, rows - 3))
-    n_buckets = len(buckets)
-    visible_bucket_count = int(_clamp(math.ceil(reveal * n_buckets), 0, n_buckets))
-    max_total = max((b.total for b in buckets), default=1) or 1
-    label_w = min(9, max(len(b.label) for b in buckets))
-    bar_w = max(14, cols - label_w - 16)
-
-    grid: Grid = []
-    labels: list[dict[str, Any]] = []
-    visible = 0
-    for i, bucket in enumerate(buckets):
-        if i >= visible_bucket_count:
-            grid.append([])
-            continue
-        visible += bucket.total
-        ink = recency_ink(bucket.rec)
-        bar_len = max(1, round((bucket.total / max_total) * bar_w)) if bucket.total else 0
-        skill_len = round((bucket.skills / bucket.total) * bar_len) if bucket.total else 0
-        if bucket.skills and skill_len == 0:
-            skill_len = 1
-        memory_len = bar_len - skill_len
-        if bucket.memories and memory_len == 0 and bar_len > 1:
-            memory_len = 1
-            skill_len = bar_len - 1
-
-        node = _bucket_label_node(bucket)
-        marker = ""
-        if node and len(labels) < 6:
-            marker = _LABEL_KEYS[len(labels)]
-            style = STYLE_MEMORY if node.get("kind") == "memory" else STYLE_SKILL
-            labels.append(
-                {
-                    "key": marker,
-                    "glyph": MEMORY_GLYPH if node.get("kind") == "memory" else SKILL_GLYPH,
-                    "label": _node_label(node),
-                    "meta": _node_meta(node),
-                    "style": style,
-                    "alpha": round(ink, 3),
-                }
-            )
-
-        cat = _bucket_category(bucket)
-        cat_hex = cmap.get(cat) if cat else None
-
-        row: Row = [[f"{bucket.label:>{label_w}} ", STYLE_LABEL, ink], ["│ ", STYLE_DIM, 0.55]]
-        if marker:
-            row.append([marker, STYLE_LABEL, 0.95])
-        elif bucket.total:
-            head_hex = cat_hex if bucket.skills else None
-            row.append(["✦" if bucket.skills else "◆", STYLE_SKILL if bucket.skills else STYLE_MEMORY, ink, head_hex])
-        if skill_len:
-            # Bar colored by the day's dominant category — a learning heatmap.
-            row.append(["━" * skill_len, STYLE_SKILL, ink, cat_hex])
-        if memory_len:
-            if memory_len == 1:
-                mem_trail = "◆"
-            else:
-                mem_trail = "◆" + ("━" * (memory_len - 2)) + "◆"
-            row.append([mem_trail, STYLE_MEMORY, max(0.65, ink)])
-        if bar_len < bar_w:
-            # Empty space keeps counts aligned; starmap texture lives in the
-            # trajectory row below, where it reads as signal rather than noise.
-            row.append([" " * (bar_w - bar_len), STYLE_BG, 1.0])
-        row.append(["  ", STYLE_BG, 1.0])
-        row.append([str(bucket.skills), STYLE_SKILL, max(0.72, ink)])
-        if bucket.memories:
-            row.append(["+", STYLE_DIM, 0.6])
-            row.append([str(bucket.memories), STYLE_MEMORY, max(0.72, ink)])
-        if i == visible_bucket_count - 1:
-            row.append(["  ◀ now", STYLE_LABEL, 0.9])
-        elif bucket.total == max_total and max_total > 1:
-            row.append(["  ☄ peak", STYLE_LABEL, 0.75])
-        grid.append(row)
-
-    # Cumulative learning trajectory underneath the rows.
-    grid.append([[(" " * (label_w + 2)), STYLE_BG, 1.0], *_trajectory_row(buckets, max(12, cols - label_w - 13), reveal)])
-
-    return {
-        "grid": grid,
-        "date": format_date(_date_at(rec, reveal)),
-        "reveal": reveal,
-        "visible": visible,
-        "labels": labels,
-    }
-
-
-# ── Trimmings ──────────────────────────────────────────────────────────────
-
-
-def build_legend(payload: dict[str, Any]) -> list[dict[str, Any]]:
-    nodes = payload.get("nodes", [])
-    skills = sum(1 for n in nodes if n.get("kind") != "memory")
-    memories = sum(1 for n in nodes if n.get("kind") == "memory")
-    return [
-        {"glyph": SKILL_GLYPH, "style": STYLE_SKILL, "label": f"skills ({skills})"},
-        {"glyph": MEMORY_GLYPH, "style": STYLE_MEMORY, "label": f"memories ({memories})"},
-    ]
-
-
-def axis_labels(payload: dict[str, Any]) -> dict[str, str]:
-    rec = compute_recency(list(payload.get("nodes", [])))
-    if not rec["timed"]:
-        return {"start": "oldest", "end": "now"}
-    return {"start": format_date(rec.get("minTs")), "end": format_date(rec.get("maxTs"))}
-
-
-def _peak_day(payload: dict[str, Any]) -> Optional[str]:
-    counts: dict[tuple[int, ...], int] = {}
-    reps: dict[tuple[int, ...], float] = {}
-    for node in payload.get("nodes", []):
-        ts = _to_ts(node.get("timestamp"))
-        if ts is None:
-            continue
-        key = _period_key(ts, "day")
-        counts[key] = counts.get(key, 0) + 1
-        reps[key] = ts
-    if not counts:
-        return None
-    best = max(counts, key=lambda k: counts[k])
-    return f"busiest day {_period_label(reps[best], 'day')} · {counts[best]} learned"
-
-
-def build_summary(payload: dict[str, Any]) -> list[str]:
-    stats = payload.get("stats", {}) or {}
-    lines: list[str] = []
-    learned = stats.get("learned_skills", stats.get("nodes", 0))
-    mem = stats.get("memory_nodes", 0)
-    edges = stats.get("related_edges", 0)
-    lines.append(f"{learned} learned skills · {mem} memories · {edges} skill links")
-    extra = []
-    if stats.get("memory_skill_edges"):
-        extra.append(f"{stats['memory_skill_edges']} memory↔skill links")
-    peak = _peak_day(payload)
-    if peak:
-        extra.append(peak)
-    if extra:
-        lines.append(" · ".join(extra))
-    return lines
-
-
-def _merge_runs(cells: Iterable[Run]) -> Row:
-    out: Row = []
-    for run in cells:
-        text, style, alpha = run[0], run[1], (run[2] if len(run) > 2 else 1.0)
-        hex_override = run[3] if len(run) > 3 else None
-        prev_hex = out[-1][3] if out and len(out[-1]) > 3 else None
-        if out and out[-1][1] == style and abs(out[-1][2] - alpha) < 1e-6 and prev_hex == hex_override:
-            out[-1][0] += text
-        else:
-            merged: Run = [text, style, alpha]
-            if hex_override:
-                merged.append(hex_override)
-            out.append(merged)
-    return out
-
-
-def render_frames(payload: dict[str, Any], *, cols: int = 80, rows: int = 16, frames: int = 48) -> dict[str, Any]:
-    """Pre-render a full play-through (reveal 0→1) plus static legend/summary."""
-    frames = max(2, min(frames, 240))
-    nodes = list(payload.get("nodes", []))
-    rec = compute_recency(nodes)
-    # Mirror render_graph's bucketing so the interactive row list lines up with
-    # what the user sees.
-    buckets = _build_chart_buckets(nodes, rec, max_rows=max(4, rows - 3)) if nodes else []
-    out_frames = []
-    for i in range(frames):
-        reveal = i / (frames - 1)
-        frame = render_graph(payload, cols=cols, rows=rows, reveal=reveal)
-        out_frames.append(
-            {
-                "reveal": frame["reveal"],
-                "date": frame["date"],
-                "visible": frame["visible"],
-                "grid": frame["grid"],
-                "labels": frame.get("labels", []),
-            }
-        )
-    return {
-        "frames": out_frames,
-        "legend": build_legend(payload),
-        "categories": category_legend(payload),
-        "buckets": _bucket_rows(buckets, payload),
-        "summary": build_summary(payload),
-        "axis": axis_labels(payload),
-        "count": len(payload.get("nodes", [])),
-        "cols": cols,
-        "rows": rows,
-    }

agent/learning_mutations.py

@@ -1,206 +0,0 @@
-"""User-initiated edit/delete for journey nodes (learned skills + memories).
-
-The journey graph (``agent.learning_graph``) gives every node a stable id:
-
-- **skills** → the skill name (e.g. ``"debugging-hermes-desktop"``)
-- **memories** → ``memory:<source>:<index>`` where ``source`` is ``memory``
-  (``MEMORY.md``) or ``profile`` (``USER.md``) and ``index`` is the node's
-  position in the combined card list (``MEMORY.md`` cards first, then
-  ``USER.md``).
-
-This module maps a node id back to its on-disk home and performs the mutation,
-shared by the CLI (``hermes journey delete|edit``), the TUI ``/journey`` overlay
-(gateway RPCs), and the desktop GUI (REST). Deleting a skill *archives* it
-(recoverable via ``hermes curator restore``); deleting a memory rewrites its
-file. Pure stdlib + existing skill/memory helpers.
-"""
-
-from __future__ import annotations
-
-from pathlib import Path
-from typing import Any
-
-_MEMORY_FILES = {"memory": "MEMORY.md", "profile": "USER.md"}
-
-
-def parse_node_kind(node_id: str) -> str:
-    return "memory" if node_id.startswith("memory:") else "skill"
-
-
-def _memories_dir() -> Path:
-    from hermes_constants import get_hermes_home
-
-    return get_hermes_home() / "memories"
-
-
-def _parse_memory_id(node_id: str) -> tuple[str, int]:
-    """``memory:<source>:<index>`` → (source, global_index)."""
-    parts = node_id.split(":", 2)
-    if len(parts) != 3 or parts[0] != "memory" or parts[1] not in _MEMORY_FILES:
-        raise ValueError(f"bad memory node id: {node_id!r}")
-    try:
-        return parts[1], int(parts[2])
-    except ValueError as exc:
-        raise ValueError(f"bad memory node id: {node_id!r}") from exc
-
-
-def _memory_local_index(source: str, global_index: int) -> int:
-    """Global card index → position within the source's own file.
-
-    ``_memory_cards`` emits all ``MEMORY.md`` cards before ``USER.md`` cards, so
-    a profile card's local index is its global index minus the memory count.
-    """
-    from agent.learning_graph import _memory_cards
-
-    cards = _memory_cards()
-    if not 0 <= global_index < len(cards):
-        raise IndexError(f"memory index {global_index} out of range")
-    if cards[global_index].get("source") != source:
-        raise ValueError("memory node id is stale — refresh the graph")
-    if source == "memory":
-        return global_index
-    return global_index - sum(1 for c in cards if c.get("source") == "memory")
-
-
-def _locate_memory(source: str, gidx: int) -> tuple[Path, list[str], int]:
-    """Resolve a memory card to its file, all §-delimited entries, and local index.
-
-    Entries come from ``MemoryStore._read_file`` — the same parser the memory
-    tool uses — so journey indices stay aligned with what the graph renders.
-    """
-    from tools.memory_tool import MemoryStore
-
-    path = _memories_dir() / _MEMORY_FILES[source]
-    if not path.exists():
-        raise ValueError(f"{path.name} not found")
-    chunks = MemoryStore._read_file(path)
-    local = _memory_local_index(source, gidx)
-    if not 0 <= local < len(chunks):
-        raise ValueError("memory node id is stale — refresh the graph")
-    return path, chunks, local
-
-
-# ── Inspect (edit prefill) ──────────────────────────────────────────────────
-
-
-def node_detail(node_id: str) -> dict[str, Any]:
-    """Current content for an edit prefill. ``content`` is the full SKILL.md
-    (skills) or the raw memory chunk (memories)."""
-    try:
-        return _node_detail(node_id)
-    except (ValueError, IndexError) as exc:
-        return {"ok": False, "message": str(exc)}
-
-
-def _node_detail(node_id: str) -> dict[str, Any]:
-    if parse_node_kind(node_id) == "memory":
-        source, gidx = _parse_memory_id(node_id)
-        _, chunks, local = _locate_memory(source, gidx)
-        body = chunks[local].strip()
-
-        return {"ok": True, "kind": "memory", "id": node_id, "label": body.splitlines()[0][:80], "content": body}
-
-    from tools.skill_manager_tool import _find_skill
-
-    found = _find_skill(node_id)
-    if not found:
-        return {"ok": False, "message": f"skill '{node_id}' not found"}
-    skill_md = Path(found["path"]) / "SKILL.md"
-    if not skill_md.exists():
-        return {"ok": False, "message": f"SKILL.md missing for '{node_id}'"}
-
-    return {
-        "ok": True,
-        "kind": "skill",
-        "id": node_id,
-        "label": node_id,
-        "content": skill_md.read_text(encoding="utf-8"),
-    }
-
-
-# ── Delete ──────────────────────────────────────────────────────────────────
-
-
-def delete_node(node_id: str) -> dict[str, Any]:
-    try:
-        return _delete_memory(node_id) if parse_node_kind(node_id) == "memory" else _delete_skill(node_id)
-    except (ValueError, IndexError) as exc:
-        return {"ok": False, "message": str(exc)}
-
-
-def _delete_skill(name: str) -> dict[str, Any]:
-    from tools import skill_usage
-
-    if skill_usage.get_record(name).get("pinned"):
-        return {"ok": False, "message": f"'{name}' is pinned — unpin it first (hermes curator unpin {name})"}
-
-    ok, message = skill_usage.archive_skill(name)
-    if ok:
-        _clear_skill_cache()
-
-    return {"ok": ok, "message": f"archived '{name}' — restore with: hermes curator restore {name}" if ok else message}
-
-
-def _delete_memory(node_id: str) -> dict[str, Any]:
-    source, gidx = _parse_memory_id(node_id)
-    path, chunks, local = _locate_memory(source, gidx)
-
-    del chunks[local]
-    _write_memory(path, chunks)
-
-    return {"ok": True, "message": f"deleted memory from {path.name}"}
-
-
-# ── Edit ────────────────────────────────────────────────────────────────────
-
-
-def edit_node(node_id: str, content: str) -> dict[str, Any]:
-    try:
-        return _edit_memory(node_id, content) if parse_node_kind(node_id) == "memory" else _edit_skill(node_id, content)
-    except (ValueError, IndexError) as exc:
-        return {"ok": False, "message": str(exc)}
-
-
-def _edit_skill(name: str, content: str) -> dict[str, Any]:
-    from tools.skill_manager_tool import _edit_skill as _do_edit
-
-    result = _do_edit(name, content)
-    if result.get("success"):
-        _clear_skill_cache()
-
-        return {"ok": True, "message": f"updated '{name}'"}
-
-    return {"ok": False, "message": result.get("error", "edit failed")}
-
-
-def _edit_memory(node_id: str, content: str) -> dict[str, Any]:
-    source, gidx = _parse_memory_id(node_id)
-    body = content.strip()
-    if not body:
-        return {"ok": False, "message": "empty memory — use delete to remove it"}
-    path, chunks, local = _locate_memory(source, gidx)
-
-    chunks[local] = body
-    _write_memory(path, chunks)
-
-    return {"ok": True, "message": f"updated memory in {path.name}"}
-
-
-# ── Helpers ─────────────────────────────────────────────────────────────────
-
-
-def _write_memory(path: Path, chunks: list[str]) -> None:
-    """Atomic temp-file + rename via the memory tool, so a concurrent reader
-    never sees a half-written file (and the §-join stays single-sourced)."""
-    from tools.memory_tool import MemoryStore
-
-    MemoryStore._write_file(path, [c.strip() for c in chunks if c.strip()])
-
-
-def _clear_skill_cache() -> None:
-    try:
-        from agent.prompt_builder import clear_skills_system_prompt_cache
-
-        clear_skills_system_prompt_cache(clear_snapshot=True)
-    except Exception:
-        pass