The .venv directory (created by uv run) contained Python 3.13 with no
deps installed. ty auto-discovers .venv for module resolution, so it
could not find discord.py, aiohttp, etc., producing ~1000 false
"Module has no member" errors.
Removing the stray .venv makes ty fall back to the nix env (Python 3.12
with all deps installed). Also set python-version = "3.12" in
[tool.ty.environment] with a comment explaining why.
ty diagnostics: 4,422 -> 3,385 (-1,037)
Tests: 496 passed, 0 failed
fix(nix): use python311 in dev shell (matches requires-python floor)
Production venv still uses python312 (nixos-unstable default). The dev
editable venv now uses python311 — the requires-python floor (>=3.11) —
so ty type checking and local dev catch 3.12+-only syntax that wouldnt
init_agent() in agent/agent_init.py sets ~176 instance attributes on the
AIAgent instance, but ty cannot track cross-module attribute assignment
through a function that receives the instance as a plain `agent` parameter.
This caused 186 unresolved-attribute errors in run_agent.py alone.
Declaring the attributes as class-level annotations (PEP 526) tells ty
they exist, clearing 182 of 190 errors (190→8). The remaining 8 are:
- 2 duck-typed object.function accesses (hasattr-guarded, safe)
- 2 missing attrs (_current_tool, _api_call_count — added)
- 2 None-narrowing on iteration_budget (.used, .max_total)
- 2 None-narrowing on client (.close)
Overall ty diagnostics: 4,648 → 4,422 (−226)
Tests: 496 passed, 0 failed
Add `| None` to ~250 params across 30 files that were annotated as bare
`str`, `list`, `dict`, `int`, `Callable`, etc. but defaulted to None.
This is a big typechecking fix. each one cascades, making ty stop
narrowing those vars to None and clearing downstream
unresolved-attribute / not-subscriptable / invalid-argument-type errors.
Two type bugs surfaced and fixed during the sweep:
1. Lowercase `callable` (builtin function) used as type annotation in
28 callback params in agent_init.py + run_agent.py. `callable | None`
isn't valid. Fixed to `Callable` (typing).
2. String forward-ref with `| None` (`"IterationBudget" | None`) is
wrong because `|` can't OR a str with NoneType. Fixed to
`Optional["IterationBudget"]`.
Also:
- Configure ty to exclude tests/ via [tool.ty.src] in pyproject.toml
(tests are ~57% of diagnostics, lowest-value typing target)
ty diagnostics: 13,290 -> 4,648 (core only, tests excluded)
Tests: 496 passed, 0 failed
Mirror CANONICAL_PROVIDERS so Fireworks sits directly under Nous Portal
(always visible) ahead of OpenRouter across onboarding, Settings → Providers,
and the API-key catalog.
* fix(credential-pool): throttle "no available entries" log to stop Windows log-lock storm
Credential selection runs on a hot path (every model call plus auxiliary
tasks), so an empty/exhausted pool logged "no available entries" at INFO on
*every* selection. On Windows, where multiple Hermes processes share one
rotating log guarded by concurrent-log-handler's cross-process lock, that
per-selection volume storms the lock (RuntimeError: Cannot acquire lock after
20 attempts), pegs a core, and stalls the asyncio event loop long enough that
the Desktop backend readiness probe times out ("Timed out connecting to Hermes
backend after 15000ms") even though the backend already announced
HERMES_BACKEND_READY.
Log the condition at most once per 60s window, re-arming on a successful
selection so recovery->re-exhaustion still surfaces promptly. Same fix class as
the warn-once dedup in #58265.
* test(credential-pool): cover no-available-entries log throttle
Assert the empty-pool INFO line logs at most once per throttle window, logs
again after the window elapses, and re-arms on a successful selection so a
recover->re-exhaust transition surfaces promptly. Uses a deterministic fake
monotonic clock (no sleeps, no network).
Complements the #65637 salvage (53d358838 + a9cc17fd8): the gateway
session store now retries transcript appends through its own queue, but
cron and CLI writers call SessionDB directly — a corrupt FTS index still
hard-failed their appends until the next process restart triggered the
offline repair.
_execute_write now detects the FTS-corruption error class (both the
generic 'database disk image is malformed' and newer SQLite's
'fts5: corrupt structure record' variant), performs a one-shot in-place
rebuild by delegating to the existing rebuild_fts(), and retries the
failed write. One-shot per instance so an unrecoverable database cannot
loop; lock/busy jitter-retry path untouched.
E2E-verified: corrupted messages_fts_data rejects appends; with this fix
the same append self-heals, persists, and FTS search works again.
Community feedback (@LSanapalli on X): the inline task-creation form is
cramped inside a ~280px column with no way to resize; board-level
workspace defaults can't be changed after board creation; and users
believe they must block a task, comment, then unblock just to talk to
a worker.
- Create-task dialog: replace the inline column form with a centered
modal (reuses hermes-kanban-dialog chrome, 36rem wide) with labeled
fields for title, assignee, priority, skills, workspace kind/path,
goal mode, and parent task. Same request shape; Enter/Escape behavior
preserved; submit disabled until a title is present.
- Board settings dialog: new Settings button in the board switcher opens
a modal to edit display name, description, and the board-level default
project directory (default_workdir). PATCH /boards/:slug now accepts
default_workdir (validated absolute existing dir; empty string clears;
omitted leaves unchanged) and returns the recomputed
default_workspace_kind so task-creation defaults follow immediately.
- Comment workflow hint: the task drawer's comment box now explains that
comments land on the thread immediately and reach the worker on its
next run/kanban_show() — no block/unblock dance needed — with a fuller
tooltip for when blocking IS the right tool.
- i18n: new keys optional in the kanban namespace with English fallbacks
in the bundle (established pattern; avoids churning 17 locale files).
- Docs: dashboard section updated for the dialog + Settings button.
A root-launched CLI session can leak /root into the terminal cwd state a
non-root gateway/cron process later resolves (#65583). os.path.isdir('/root')
is True for a non-root user — stat only needs search permission on / — so
_resolve_safe_cwd returned it and subprocess.Popen(cwd='/root') died with
PermissionError: [Errno 13], failing EVERY cron job's terminal/file/search
tool on every command until restart.
_resolve_safe_cwd now requires X_OK (new _cwd_usable helper) and climbs to
the nearest enterable ancestor, logging a WARNING that names the leak class
when an existing-but-denied cwd is skipped. Missing-cwd recovery (#17558)
behavior unchanged.
E2E-verified: LocalEnvironment constructed with an unenterable cwd now runs
commands from the fallback directory instead of raising.
The widened resume sites pass repair_alternation=True; the DB doubles in
tests/test_tui_gateway_server.py (separate from tests/tui_gateway/) needed
the same signature update as Frowtek's originals.
Sibling-site audit on top of #65672: the interactive TUI resume, the
profile-scoped resume, and the /undo history reload also feed LIVE
REPLAY (raw_history -> sanitize_replay_history -> working conversation;
session['history'] after rewind). Pass repair_alternation=True on the
model-fed copies; display_history stays verbatim so inspection/export
show what is actually stored. Display-only consumers (session.history
RPC, formatted transcript output) intentionally unchanged.
test_build_child_agent_ignores_acp_command_when_binary_missing patches
shutil.which globally and asserted the LAST call was which("copilot").
That is order-dependent: an unrelated which("uv") reached later in the same
process (which happens under some CI test-slice orderings) becomes the last
call, so assert_called_with("copilot") fails even though the copilot binary
was probed exactly as intended. Switch to assert_any_call("copilot"), which
verifies the actual intent and is robust to unrelated which() calls. The
behavioural assertions (provider, acp_command, acp_args) are unchanged.
The lazy session.resume path now calls
db.get_messages_as_conversation(target, repair_alternation=True), but the
fake _DB stubs in test_protocol.py still declared the pre-change signature,
so the resume raised "unexpected keyword argument 'repair_alternation'"
and the three session_resume_lazy tests failed.
Mirror the real get_messages_as_conversation signature in the stubs by
accepting (and ignoring) repair_alternation.
Follow-up to the restore-boundary alternation heal (#65492): get_messages_
as_conversation grew a repair_alternation flag, wired into gateway
load_transcript and the CLI startup resume. Three other LIVE-REPLAY
restore sites still loaded the transcript verbatim, so a durable
'user;user' violation there re-fires the pre-request defensive repair on
every request for the rest of the session (it only ever mutates the
per-request list, never the restored working conversation):
- acp_adapter/session.py::SessionManager._restore — the loaded history
becomes the resumed ACP (Zed) agent's SessionState.history.
- hermes_cli/cli_commands_mixin.py — the /resume slash command sets
self.conversation_history from the load (the startup resume was fixed,
this mid-session one was missed).
- tui_gateway/server.py — the resume handler feeds the load into the
deferred session record's working conversation.
Pass repair_alternation=True at all three so the wedge is healed once at
restore. Inspection/export consumers (trace upload, context guard,
api_server history, display_history) keep the verbatim default.
Adds an end-to-end regression test driving the ACP _restore path: a
seeded user;user session restores to an alternation-clean live history
with no user input lost.
Two more display gaps from #26541 grafted onto the merged bridge:
- webSearch: codex's built-in web search now produces a tool.started/
tool.completed bubble pair (query as preview + args). Previously the
item type wasn't in _CODEX_TOOL_ITEM_TYPES, so built-in searches
showed nothing.
- mcp.hermes-tools.* stripping: tools codex invokes through Hermes' own
hermes-tools MCP server display as their bare names (web_search,
browser_navigate) instead of mcp.hermes-tools.web_search. The inner
dispatch subprocess can't fire native progress events, so the
codex-level event is the display event — name it the way users know
the tool.
Credit: both behaviors designed and first implemented by @simpolism in
PR #26541 (May 15, earliest of the app-server display-bridge family).
The approval-drain loop in CodexAppServerSession.run_turn drains up to 8
pending notifications to keep per-turn state current before answering a
server-initiated approval request — but never forwarded them to the
on_event display hook. Tool bubbles for items drained alongside an
approval (e.g. the item/started for the very command awaiting approval)
silently disappeared.
Mirror the main notification path's on_event invocation in the drain
loop. Regression test demonstrates RED→GREEN.
Grafted from PR #26541 by @simpolism — the earliest submission of the
codex app-server display-bridge family (May 15). Confirmed independently
by #64698 and #65412.
The REFERENCE ONLY framing ('treat as background reference, NOT as active
instructions... Do NOT answer questions or fulfill requests') was observed
bleeding into general tool-use suppression: a production session went
narration-only for 7 consecutive turns immediately after a compression
event, describing edits instead of calling tools (#65848 report).
Fix is additive: one clause stating the note does not restrict HOW the
agent works — tools remain fully active for the active task. Every
anti-resumption protection stays intact; the previous prefix generation
is frozen into _HISTORICAL_SUMMARY_PREFIXES per the module contract so
persisted summaries still get the directive-strip on re-compaction.
The #65848 rewrite was not taken: dropping the 'Do NOT answer questions'
line and the four-heading discard directive risks re-opening the
stale-task-resumption class those clauses exist to prevent (the carveout
era regressions #41607/#38364/#42812 documented in this file).
Report and root-cause analysis: @yasserbousrih (#65848).
Widen the #65991 single-writer fence to run_codex_stream: each codex
attempt claims the delta sink before consuming events, and the consume
loop's interrupt_check now also stops the instant a newer attempt
supersedes this one. Parity with the chat_completions / anthropic /
bedrock paths from the salvaged fix.
Two regression tests: superseded codex stream is fenced mid-stream;
sole-writer codex stream delivers unchanged.
Assert that a superseded stream (older writer token, other thread) is fenced
from the delta sink, the active writer is never fenced, a non-claiming thread
is never treated as a writer, and the real consume loop stops the instant it is
superseded — so two streams can never interleave into one turn (#65991).
When the stale-stream detector reconnects past a stream whose socket abort
raced (the close never actually stopped the old worker), the superseded stream
and the retry's stream both write deltas into the same turn. The persisted
transcript is then two coherent responses interleaved token-by-token —
de-interleaving the stored text by alternation yields two complete, independent
answers to the same prompt, which is a dual-writer race in the harness, not a
model/context failure (#65991).
The interrupt path already positively cancels before force-closing (#6600), but
the stale-kill path relies only on the socket abort, and nothing fenced late
chunks from a superseded stream out of the shared delta sink.
Enforce a single-writer invariant on the sink itself, guarded by attempt id
rather than only socket state: every streaming attempt (chat_completions,
anthropic_messages, and bedrock paths) claims a monotonic writer token before
it begins consuming its stream. A newer claim supersedes any older one, so the
consume loop bails the instant it is superseded and _fire_stream_delta /
_fire_reasoning_delta / _record_streamed_assistant_text drop chunks from a
stale writer. The token is stored per-thread, so a thread that never claimed
(a non-streaming delta caller) is never fenced — the guard can only ever drop a
superseded stream, never the single legitimate writer. Discards are counted and
logged sparsely so a real provider problem stays visible instead of being
silently swallowed.
A hermes --tui session whose main thread wedges before app.run() returns
never executes the finally that calls _run_cleanup — the only place the
exit watchdog was armed — so a dead CLI lingered indefinitely (observed
~47 min at 4% CPU, the #65998 class).
Arm the backstop from the SIGTERM/SIGHUP handlers instead (both the
interactive and single-query paths), the earliest moment shutdown intent
is unambiguous. The signal-armed leash is 2x HERMES_EXIT_WATCHDOG_S so a
slow-but-progressing _run_cleanup (which still arms its own tighter timer)
is never cut short; the outer timer only wins when cleanup was never
reached. Idempotent across repeated signals; never raises from a handler.
Deliberately NOT armed at startup: the watchdog thread calls os._exit(0)
unconditionally after its sleep, so a startup-armed timer (the #65998
approach) would hard-kill every session that outlives the timeout.
Supersedes #65998; thanks @JeffStone69 for the report and root-cause gap
analysis.
Degrading models (observed with gpt-5.6 past ~350K input) emit tool-call
batches like 6 valid named calls + 1 blank-name call. Previously the
whole turn was voided — every valid call got 'Skipped: another tool call
in this turn used an invalid name' — and three such batches tripped the
3-strike stop, killing sessions that were still making progress.
Now a mixed batch error-results ONLY the invalid call(s) (terse
anti-priming error for blank names per #47967, catalog dump for typos)
and dispatches the valid subset for execution. The assistant message
keeps every emitted call so provider-side tool_call/result pairing stays
intact. The 3-strike counter only advances when a turn contains NO valid
call, so a fully-degenerate model still stops while a mostly-coherent
one keeps working. Broken JSON args on a never-executing invalid call no
longer trigger the whole-turn JSON retry loop.
Field evidence: July 2026 debug bundle showed gpt-5.6-sol emitting
6-call batches with one blank-name rider at 559K/384K-token context in
two separate sessions; 13 valid tool calls were discarded before the
session stopped as partial.
Follow-up fixes for salvaged PR #65637:
1. Clear _dirty_transcripts in rewrite_transcript + rewind_session —
stale pending messages were re-inserted after /retry, /undo, /compress.
2. Narrow _is_fts_corruption_error to specific SQLite error strings —
bare 'fts' substring matched 'shifts', 'gifts', etc.
3. Move DB write outside _transcript_retry_lock — holding the lock
during writes serialized all sessions' transcript appends and blocked
during FTS rebuild. Now the lock guards only the pending queue.
4. Push rebuild_fts() into SessionDB — SessionStore was reaching into
_conn/_lock private attrs. SessionDB.rebuild_fts() follows the same
pattern as optimize_fts().
5. Cap pending per session at 200 — prevents unbounded memory growth
when DB is persistently broken. Oldest messages dropped with warning.
Added 4 new tests: dirty-clear on rewrite/rewind, FTS matcher false
positives, pending cap enforcement.
Two changes to tests/agent/test_file_safety_session_state.py:
1. Drop the stale monkeypatch on tools.file_tools._get_live_tracking_cwd
— the helper was deleted in the cwd-tracking refactor (c80b244b5),
and monkeypatch.setattr on a missing attribute raises AttributeError,
breaking CI slice 4/8 on main for every PR. The patch was redundant:
the test writes an absolute path, so cwd resolution never engages.
2. Make the fixture stale-proof: instead of monkeypatching the private
_hermes_home_path/_hermes_root_path helpers (same failure class if
they're ever renamed), set HERMES_HOME to <root>/profiles/work and
let the real resolution chain (get_hermes_home /
get_default_hermes_root's profiles-parent rule) derive both paths.
The fixture now references zero private symbols and exercises the
production resolution path.
Port from anomalyco/opencode#35671: Z.AI / Zhipu GLM returns
'tokens in request more than max tokens allowed' (error code 1210) on
context overflow. This matched no pattern in _CONTEXT_OVERFLOW_PATTERNS,
so the error classified as unknown/retryable — the agent would retry the
oversized request instead of triggering context compression.
Proven live on main before the fix: classify_api_error() returned
FailoverReason.unknown for the exact Z.AI error shape; now returns
context_overflow.
Per the MCP spec the cursor is an opaque string; anything else
(including MagicMock auto-attributes in tests) means no more pages.
Fixes test_mcp_tool_session_expired mock-session runaway.
Port from anomalyco/opencode#35439/#35500: preserve full MCP catalogs
across paginated tools/list responses.
The MCP spec allows servers to paginate tools/list, resources/list, and
prompts/list via an opaque nextCursor token. The Python SDK's
ClientSession.list_* methods fetch exactly one page per call, and hermes
never passed the cursor back — on a paginated server every tool,
resource, and prompt past page 1 was silently invisible to the agent.
Adds _paginate_full_list() (cursor-draining helper with a 50-page
runaway cap and spec-correct opaque-string cursor validation) and
applies it at all three discovery sites: _discover_tools(), the
tools/list_changed refresh handler, and the list_resources/list_prompts
utility handlers. The keepalive probe intentionally keeps its
single-page call (liveness only).
E2E: real stdio MCP server serving 3 pages (2/1/1 tools) — old code
discovered 2 tools, new code discovers all 4.
Port from Kilo-Org/kilocode#11955: Gemini validates every object schema's
required list strictly against the same node's properties and fails the
ENTIRE GenerateContentRequest with HTTP 400 'required[0]: property is not
defined' when a name has no matching property. MCP servers (e.g. the GitHub
remote MCP) routinely emit array item schemas carrying required without
properties, which made every request on the native Gemini path fail before
any model output.
sanitize_gemini_schema() now filters required to names present in the node's
properties and drops the keyword when nothing valid remains. Applies
recursively (properties / items / anyOf). Tool handlers still validate
required fields at execution time, so nothing the model could actually use
is lost.
Scoped to the Gemini-facing sanitizer only — the universal
tools/schema_sanitizer.py already prunes typed object nodes, and its
remaining gap (untyped nodes) is contested by open PR #20151.
Port from openai/codex#31494: user-visible history replay must strip CSI
sequences and control characters. Stored conversation history can carry
raw terminal escapes (pasted content, gateway-origin text, model output
echoing injected tool results). Replaying it via /resume's recap panel or
build_recap (/status on CLI + gateway) wrote those bytes straight to the
terminal — an injected message could clear the screen, retitle the window,
move the cursor, or restyle the recap UI. Rich's Text() does not neutralize
raw escape bytes.
- tools/ansi_strip.py: add sanitize_display_text() — strip_ansi() plus
bare C0/C1 control removal, preserving \n and \t, normalizing \r to \n
(adapted to Python from Codex's sanitize_user_text; reuses the existing
ECMA-48 stripper instead of transcribing their char-walk)
- hermes_cli/cli_agent_setup_mixin.py: sanitize user + assistant text in
_display_resumed_history() before building the Rich recap panel
- hermes_cli/session_recap.py: sanitize preview lines in build_recap()
(_truncate choke point) so /status recaps are clean on every platform
- tests: 10 new sanitize_display_text cases (incl. the exact codex#31494
fixture), recap + resume-display leak assertions
The previous commit fixes the canonical agent/skill_utils.parse_frontmatter.
Six more modules reimplement the '---' fence check locally and had the
same bug:
- tools/skill_manager_tool.py _validate_frontmatter — rejected BOM'd
skill_manage create/edit content outright
- tools/skills_hub.py GitHubSource._parse_frontmatter_quick and
OptionalSkillSource._parse_frontmatter — hub browse/install metadata
- hermes_cli/skills_hub.py — local skill install validation
- gateway/run.py — skill slug discovery for disabled-skill hints
- agent/prompt_builder.py _strip_yaml_frontmatter — BOM'd context files
(AGENTS.md) leaked raw frontmatter into the system prompt
- tools/blueprints.py _split_frontmatter — str.lstrip() does not strip
U+FEFF (not whitespace), so the existing lstrip never covered it
Sibling-surface regression tests added.
Bug class also fixed upstream in cline/cline#12218 (found by the weekly
Cline PR scout).
A UTF-8 BOM saved into a SKILL.md (e.g. Notepad or PowerShell `>`) is kept by
read_text(encoding="utf-8"), so the string handed to parse_frontmatter starts
with the BOM and the startswith("---") fence check fails. The whole frontmatter
is then silently dropped: the skill loads with no name/description, `platforms`
gating falls open (a macOS-only skill becomes visible everywhere), and
required_environment_variables / metadata.hermes.config setup never fires.
Strip a single leading BOM at the top of parse_frontmatter, the shared
chokepoint for every local skill-loading path (_parse_skill_file,
discover_all_skill_config_vars, DESCRIPTION.md parsing, _inject_skill_config,
and the tools/skills_tool._parse_frontmatter re-export), so the whole class is
covered, not just the reported site. Only the leading marker is removed; a BOM
mid-content is left as data. Mirrors the existing file-tools BOM handling
(#35278) and CONTRIBUTING.md "File encoding".
Adds tests: BOM'd frontmatter parses identically to plain, the body is
BOM-free, platforms gating and config-var extraction survive a BOM, and an
end-to-end BOM-write / plain-read round trip (mirroring _parse_skill_file).
test_auto_reset_does_not_recover_session_being_ended asserted the old
end_session('session_reset') call; auto-reset now writes through
promote_to_session_reset with the specific auditable reason
('suspended' for a suspended entry). Missed in the local targeted run
because the file wasn't in the touched-suite list; caught by CI shard 6.
Unifies the two gateway subsystems that were fighting each other: the
'never lose a session' recovery machinery (#54878 stale-route self-heal,
find_latest_gateway_session_for_peer reopening agent_close/ws_orphan_reap
rows) and the session reset/expiry machinery (expiry watcher, /new,
/resume, resume_pending freshness gate).
The unified contract:
- INTENTIONAL boundaries (expiry finalization, auto-reset, /new,
/resume switch) are recorded durably via promote_to_session_reset(),
which upgrades accidental recoverable end_reasons (agent_close,
ws_orphan_reap) to the explicit boundary while preserving other
explicit reasons (compression, etc.). Recovery then correctly refuses
to resurrect them.
- ACCIDENTAL ends (crash, cleanup bug, mistaken reaper) stay
recoverable — genuine crash recovery is untouched.
On top of the cherry-picked contributor commits:
- promote_to_session_reset widened to ws_orphan_reap + parameterized
reason so auto-reset paths stay auditable (idle/daily/suspended/
resume_pending_expired) (#61220, #61993, #63539)
- get_or_create_session auto-reset, reset_session (/new), and
switch_session (/resume) all write through the promote path — the
first-reason-wins end_session no-op could previously leave a reset
session resurrectable behind a stale agent_close row (#61993)
- resume_pending freshness gate now honors session_reset.mode=none:
explicit opt-out of automatic resets also opts out of the zombie
gate (#61052)
- resume recovery note extracted to build_resume_recovery_note() and
made adapter-aware via a new interactive_resume capability flag:
webhook/api_server auto-resume turns now CONTINUE the interrupted
task instead of emitting an unanswerable 'session restored'
acknowledgement that abandoned the work (#57056)
- tests updated to call the real note builder instead of mirroring it
E2E-validated against a real SessionDB + SessionStore in a temp
HERMES_HOME: expiry->agent_close->no-resurrection, /new promote,
crash recovery preserved, mode=none opt-out, routing-table flag sync.
When a gateway session with resume_pending=True is not recovered within the
auto-continue freshness window (e.g. because repeated API calls timed out on a
large context), get_or_create_session correctly creates a new session. However
two gaps existed:
1. The user received no notification — resume_pending_expired fell through the
generic "inactive for Xh" else-branch in run.py, which produces wrong wording
and (for session_reset.mode: none users) is gated on policy.notify that
evaluates to False.
2. The old session was ended in state.db with the hardcoded generic reason
"session_reset", making it impossible to distinguish from a normal idle/daily
reset in post-mortem analysis.
Fix:
- gateway/run.py: add an explicit resume_pending_expired case for the agent
context note ("gateway restart recovery timed out") and the user-facing
notification. Always notify for this reason — like suspended — because the
user had an active session that was silently replaced.
- gateway/session.py: pass auto_reset_reason as the DB end_reason instead of
the hardcoded "session_reset", so all auto-reset paths are auditable.
- tests: extend TestResumePendingExpiredAutoReset in test_session_reset_notify.py
with five new cases that cover the reason, activity flag, DB end_reason,
non-regression of the idle path, and freshness-disabled bypass.
Closes#58933
Co-authored-by: Cursor <cursoragent@cursor.com>
When the #54878 self-healing path drops a stale sessions.json entry, the
fix at gateway/session.py:1765 now checks _should_reset() before falling
through to DB recovery. This test covers the case where the stale entry's
session is overdue under an idle/daily reset policy — it must create a
fresh session, set auto-reset metadata, and NOT call reopen_session().
When the session expiry watcher finalizes a session (daily/idle reset)
and the next message triggers the #54878 self-healing path
(get_or_create_session detects sessions.json / state.db mismatch),
the recovered session was silently reopened without checking whether
it should have been reset. This caused sessions to persist indefinitely
across reset boundaries.
Fix: after dropping the stale sessions.json entry in the self-healing
path, call _should_reset() against the old entry's updated_at. If a
reset is due, set db_end_session_id to skip DB recovery and create a
fresh session — matching the normal reset flow.
Address review feedback on #63068:
1. Replace unconditional reopen_session() + end_session() with a
conditional promote_to_session_reset() method in SessionDB.
The new method only promotes live rows or rows ended with
agent_close — explicit boundaries (compression, session_reset,
new_command) are preserved via first-writer-wins semantics.
2. Rewrite tests to use real SessionDB instead of MagicMock:
- 7 unit tests for promote_to_session_reset edge cases
- 3 integration tests verifying the actual recovery contract
in find_latest_gateway_session_for_peer after promotion
Port from anomalyco/opencode#36130: the Responses spec carries streaming
error details at the top level of the error frame, but the official OpenAI
SDK and several OpenAI-compatible proxies wrap them in an HTTP-style nested
envelope ({"type": "error", "error": {code, message, param}}).
_raise_stream_error only read top-level fields, so nested-envelope frames
collapsed to the generic 'stream emitted error event' placeholder with
code=None — the error classifier never saw the provider's real failure
reason, misrouting rate-limit / context-overflow / entitlement errors into
the generic retry path.
Top-level fields keep precedence; the envelope is a fallback. Null-tolerant
for spec-compliant frames with explicit nulls.
Only advertise finite watchdog deadlines that are still in the future, exercise the full MoA heartbeat path, and register the salvaged contributor attribution.
The dashboard console previously ran under a 'hosted' context that
blocked most commands (auth add, config set model.*, mcp add --command,
cron --script, ...) behind an allowlist + line-policy layer. With the
full Hermes CLI now built into the dashboard, that policy layer is
redundant gatekeeping: the console gets the same command surface
everywhere.
Removed:
- ConsoleContext/contexts plumbing on ConsoleCommand + engine
- EXPECTED_HOSTED_PATHS allowlist + _mark_hosted
- _enforce_hosted_line_policy + HOSTED_CONFIG_* allow/block tables
- _dashboard_console_context() and the context field on the ready frame
- hosted-context tests; context badge in HermesConsoleModal
Kept (mechanical, not policy): shell-syntax rejection, the
interactive/server command blocks (gateway, dashboard, mcp serve, ...),
mutating-command confirmations, output caps, and command timeouts.
Follow-ups on top of @xxxigm's salvaged bridge (#33294):
- Remove the now-dead narrow item/started-only mapper from #38835
(_codex_note_to_tool_progress) — the full bridge supersedes it and
keeps the same tool-name contract; its tests are repointed at the
bridge helpers.
- Preserve main's request_routing/approval-bypass wiring on the
CodexAppServerSession constructor (landed after the PR was filed).
- Gate agentMessage interim delivery on display.show_commentary so the
app-server runtime honors the same toggle as the codex_responses
commentary channel (tool progress is unaffected).
- Add json import (bridge helpers use json.dumps) and modernize the
wiring test's stub agent for main's usage-accounting attributes.
42 tests across five suites:
* ``TestCodexItemToToolName`` / ``TestCodexItemToArgs`` /
``TestCodexItemToPreview`` / ``TestCodexItemCompletionPayload`` —
pin the per-type mapping so the synthetic tool name + args the
UI sees match what ``CodexEventProjector`` writes into messages.
* ``TestStreamDeltaDispatch`` / ``TestToolProgressDispatch`` /
``TestAgentMessageInterimDispatch`` — drive each Codex
notification shape through the bridge and assert the right
agent callback fires with the right arguments (including the
duration / is_error / result kwargs the gateway renders).
* ``TestBridgeRobustness`` — defensive paths: non-dict
notifications, missing params, raising callbacks (must not
tear down the codex turn loop), and agents without callbacks
registered (cron / gateway-less contexts).
* ``TestBridgeWiredInRuntime`` — integration guard that
``run_codex_app_server_turn`` actually constructs the session
with ``on_event=<bridge>``, preventing a future refactor from
silently regressing live progress visibility again.
Pass ``on_event=make_codex_app_server_event_bridge(agent)`` when
spawning the per-session ``CodexAppServerSession``. The session has
always had a raw event hook but ``run_codex_app_server_turn`` never
supplied one, so Discord / Telegram / TUI users saw nothing while
codex was working — only the final answer landed.
Now each ``item/started`` for a tool-shaped item fires
``tool_progress_callback("tool.started", ...)``, ``item/completed``
fires the matching ``"tool.completed"`` with duration + result,
``item/agentMessage/delta`` flows through ``_fire_stream_delta`` and
each completed ``agentMessage`` surfaces through
``_emit_interim_assistant_message`` so the gateway's
``already_streamed`` dedupe keeps interim commentary in the channel
without duplicating text the stream already showed.
Adds ``make_codex_app_server_event_bridge(agent)`` plus four small
mapping helpers (``_codex_item_to_tool_name`` / ``_codex_item_to_args``
/ ``_codex_item_to_preview`` / ``_codex_item_completion_payload``)
that translate codex JSON-RPC ``item/*`` notifications into the
exact shape Hermes' gateway UI callbacks expect — tool names match
``CodexEventProjector`` so the progress bubbles and the projected
``tool_calls`` entries use the same identifiers.
No behaviour change yet: the next commit wires the bridge into
``run_codex_app_server_turn`` (#33200).
Review follow-up: the store-internals tests proved _current_cron_store()
resolves lazily, but not that the PUBLIC job I/O honors it. This exercises
save_jobs()/load_jobs() after a late env repoint and asserts the
import-time jobs.json stays byte-identical to a planted sentinel.
Review follow-up: tests that monkeypatch CRON_DIR/JOBS_FILE/OUTPUT_DIR (the
documented process-wide compatibility surface) were bypassed by the lazy
env fallback — 3 file-permission tests, the cross-process lock test, and
the heartbeat roundtrip regressed. _current_cron_store() now snapshots the
constants at import and honors any deliberate re-point of them ahead of
the env resolution, so the precedence is: use_cron_store() override >
patched constants > fresh HERMES_HOME > import defaults. Adds a test
pinning constants-beat-env; the late-env sentinel behavior is unchanged.
tests/cron: failure set byte-identical to unpatched main on this box
(the 5 regressions gone); 138 pass in the touched files.
Complements ec0227b43 (context-scoped cron store): the ContextVar override
is the right tool for deliberate cross-profile scoping, but with no
override active, _current_cron_store() returned the import-time constants —
so a HERMES_HOME set AFTER cron.jobs import (the filed incident: test
fixtures patching the env too late) still read/wrote the user's real
jobs.json. The fallback now resolves the active profile home fresh via
get_hermes_home() (context-local override, then env) and scopes the store
to it; when the home is unchanged since import, the exact module-level
constants are returned as before (zero change in the common path, and they
remain the documented compatibility surface). use_cron_store() still wins.
Three tests: late env repoint scopes the store; unchanged home returns the
import-time constants identically; an active use_cron_store() override
beats the env.
* feat(browser): store full snapshots on truncation; make eval denylist opt-in
Two harness fixes motivated by BU_Bench results where fixed-verb + lossy
observation cost Hermes heavily vs code-driven browser agents:
1. Snapshot truncation no longer loses content. When a snapshot exceeds
the 8000-char threshold, the complete accessibility tree is saved to
cache/web (same truncate-and-store pattern as web_extract) and the
truncated view / LLM summary includes the file path plus a ready-made
read_file call. Element refs beyond the cut are recoverable without
re-snapshotting. Stored copies are force-redacted and capped at 2MB;
content-hash filenames dedupe repeated snapshots of the same page.
2. The browser_console(expression=...) sensitive-primitive denylist is
now opt-in via browser.restrict_evaluate (default false). The
names-based denylist blocked legitimate DOM extraction — any selector
or expression containing 'fetch', 'cookie', 'input', etc. — which
crippled the agent's only programmatic page-inspection path. The
SSRF/private-URL egress guards in _browser_eval are independent of
this policy and remain always-on. browser.allow_unsafe_evaluate keeps
its meaning (bypass the denylist) for configs that already set it.
* test: update None-guard test for stored-snapshot pointer in _extract_relevant_content
test_normal_content_returned pinned the exact return value; the summary
now carries a pointer to the stored full snapshot. Assert the summary
passes through and the pointer is present instead.
* feat(browser): align snapshot threshold with web_extract's 15k char budget
SNAPSHOT_SUMMARIZE_THRESHOLD 8000 -> 15000, matching
web_tools.DEFAULT_EXTRACT_CHAR_LIMIT so the snapshot and web_extract
truncate-and-store paths give the model the same per-page budget.
_truncate_snapshot's default max_chars now follows the constant.
Invariant test added; docs (en+zh) and CLI tip updated.
Commentary delivery is on by default; users who find the extra mid-turn
narration noisy can set display.show_commentary: false to restore the
previous behavior (commentary routed to the reasoning channel, visible
only with show_reasoning).
- hermes_cli/config.py: display.show_commentary default true
- agent/agent_init.py: wire config -> agent.show_commentary
- run_agent.py: gate structured commentary extraction on the flag
- agent/codex_runtime.py: gate live-stream commentary callback (falls
back to legacy reasoning-channel routing when off)
- docs + 2 tests (interim path off, live stream fallback)
Also adds AUTHOR_MAP entries for davidrobertson and 100yenadmin.
The openai-codex and xai-oauth branches of _refresh_entry duplicated the
lock-timeout computation and _auth_store_lock acquisition. Extract the
shared scaffolding: a combined provider guard, a dispatch to the
provider-specific sync helper, and a _single_use_refresh_lock_timeout()
helper. Each provider's distinct post-sync decision logic (codex
needs-refresh short-circuit vs xai token-equality adoption) is preserved
verbatim. Behavior parity verified by the credential pool suite (98
passed) and a direct timeout-helper probe for both providers.
Follow-up to salvaged PR #62285.
Keep each xAI OAuth auth-add login as an independent manual device-code pool entry and recognize xAI personal-team spending-limit 403 responses as billing exhaustion. Preserve the structured top-level error message so the failed credential is quarantined and the next healthy account is selected without attempting a pointless token refresh.
Route direct xAI HTTP consumers through the credential pool as well. Proactive and 401-reactive refreshes update the exact issuing manual entry, preserve validated xAI base URL overrides, and serialize single-use refresh-token rotation across concurrent pool instances.
Add a runtime_validator callback to generate_title() / auto_title_session()
/ maybe_auto_title(). Callers snapshot the session's model+provider when
spawning the background titler; the validator runs right before the LLM
request and skips it silently when the live runtime no longer matches —
so a stale title request can't reload a model that strict_single_load
already evicted after a user model switch. Fail-open: a raising validator
never disables titling.
Wired at all four call sites (cli, gateway, tui_gateway, acp_adapter).
Surgical reapply of PR #19137 (base was 8k+ commits stale; the original
patch predates the pinned-language prompts, the atomic-write helper, and
the moved TUI/ACP call sites). Original work by @Thatgfsj. Closes#19027.
Combines the two salvaged fixes so they compose instead of conflict:
_persist_session_title (#50575) now writes through set_auto_title_if_empty
(#51483) when the store provides it — the collision-dedup retry and the
manual-/title race protection apply together. Predicate failure (a manual
title landed while generation was in flight) returns None: nothing written,
no callback. Legacy stores without the atomic method keep the plain
set_session_title path, including the vanished-session RuntimeError.
Tests cover both store shapes plus the race-skip path; E2E verified against
a real SQLite SessionDB (collision -> 'Weekly Report #2', manual title
preserved, cron dedup, blank guard). AUTHOR_MAP entry for rasitakyol.
When the terminal's own 'disable mouse reporting' toggle (or an external
app / tmux) clears the DEC mouse modes, a mouse-only user is deadlocked:
every existing recovery trigger (resize, >5s stdin gap + keypress,
raw-mode bounce) needs stdin — and mouse reporting being off is exactly
why no stdin arrives. Users had to resize the window to scroll again.
Fix: a mouse-mode watchdog in App that DECRQM-probes mode 1000 every 2s
while tracking is expected on. If the terminal reports the mode RESET,
reassertTerminalModes() re-arms tracking — the same recovery a resize
performs, without the resize. Probes are skipped whenever a mouse/wheel
event arrived within the interval (tracking provably alive → zero query
chatter during normal use), never fired while paused for an editor
handoff or when /mouse off was chosen, and the watchdog permanently
disables itself on terminals that don't answer DECRQM (DA1-sentinel
resolution via the existing timeout-free TerminalQuerier).
Terminals whose toggle merely gates event delivery report SET, so an
active user toggle is never fought; terminals whose toggle clears the
modes report RESET after re-enable and recover within ~2s.
Follow-ups for the consolidated salvage:
- Memoize the config.yaml-derived timeout on the file's mtime_ns so the
rebuild-on-timeout-change check from PR #57437 costs one stat() per
get_honcho_client() call instead of a full YAML load on the hot path.
- hermes honcho status now displays the host-block-resolved
dialecticCadence (remnant from PR #63776, whose runtime fix landed
in #62290).
- AUTHOR_MAP entries for the salvaged contributor emails.
The Honcho client singleton cached the HTTP timeout at first build.
In long-lived processes (gateway, dashboard), changing the timeout
via config.yaml or HONCHO_TIMEOUT had no effect until restart.
Track the resolved timeout alongside the cached client and compare
on each get_honcho_client() call. When the timeout differs, reset
the singleton so the next call rebuilds with the new value.
Fixes#57347
Fixes Honcho client initialization for setup-generated configs that store
connection details in a named host block (e.g. "local"). Previously:
- base_url was only read from flat config root, not from host_block.
- resolve_active_host() ignored defaultHost and always used the Hermes
profile key ("hermes"), so the host block lookup returned {} and the
api_key was also lost.
FixesNousResearch/hermes-agent#61661
honcho_reasoning's minimal tier hard-caps Honcho's dialectic output at 250
tokens combined with the model's own hidden reasoning tokens. Confirmed via
direct honcho-api server logs that a multi-fact query ("summarize known
facts about this peer and communication preferences") run at
reasoning_level=minimal gets cut off mid chain-of-thought at exactly
output_tokens=250, before the model ever reaches a synthesized answer.
low/medium/high/max fall back to Honcho's much larger global dialectic
default and don't hit this cap. Since dialecticDynamic is on by default and
the calling model picks reasoning_level itself via this tool parameter, the
fix is to make the tradeoff explicit in the parameter description so the
model defaults to low unless the query is genuinely a single-fact lookup.
Schema-description-only change; the shared dialectic_max_chars truncation
cap in session.py's dialectic_query() is a separate fix (its own PR).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Verify that get_prefetch_context queries user context through the
assistant observer when _ai_observe_others is enabled, matching
the fix that routes _fetch_session_context through
_resolve_observer_target.
_fetch_session_context was using user_peer_id directly as observer
when calling _fetch_peer_context, bypassing _resolve_observer_target.
This caused honcho_context to return empty data because the observer
perspective was wrong (user instead of hermes/assistant).
Fixed by resolving observer via _resolve_observer_target(session, 'user'),
consistent with all other call sites (get_peer_card, honcho_search, etc.).
- Make the config imports lazy inside _auto_title_enabled(), matching the
existing _title_language() pattern (title_generator is imported from agent
code paths where a module-level hermes_cli import risks circularity).
- Check the enabled flag after the cheap first-exchange guard in
maybe_auto_title so config isn't read on every turn of a long session.
- Repoint the two new tests at the real import site.
- Document the key in cli-config.yaml.example and merge the enabled flag
into the existing title_generation block in configuration.md.
- AUTHOR_MAP entry for the contributor.
* fix: detect env-var-configured providers absent from PROVIDER_REGISTRY
is_provider_explicitly_configured() only checked PROVIDER_REGISTRY (a
manually-maintained dict) for env-var names. Providers that exist solely
in the models.dev catalog — e.g. openrouter — were never recognised as
explicitly configured, so they were filtered out of the desktop model
picker even when their API key was set in .env.
Add a fallback to get_provider() (which reads the models.dev catalog)
when PROVIDER_REGISTRY returns None. Both ProviderConfig and ProviderDef
expose .auth_type and .api_key_env_vars with the same shape.
* test: keep OpenRouter provider gate assertion behavioral
* chore(release): map salvaged OpenRouter contributor
---------
Co-authored-by: zzpigpinggai <zzpigpinggai@users.noreply.github.com>
Add a developer-guide page for running the Ink TUI and Electron desktop
app from a git worktree without a full npm install per checkout, via the
htui/hgui shell helpers that share node_modules from a canonical deps
checkout by symlink (falling back to a local npm ci when the lockfile
diverges). Registers it in the sidebar, cross-links from the TUI and git
-worktrees pages, and documents the previously-undocumented
HERMES_DESKTOP_PYTHON / HERMES_DESKTOP_DEV_SERVER env vars the desktop
backend reads.
Follow-up to #65890 (router transitions off) and #65898 (structural
compare + first-paint budget): profiling the switch path on real 1000+-
message sessions with a new CDP harness showed the remaining freeze is
NOT markdown rendering — it's a forced-reflow cascade from mount-time
layout reads interleaved with style writes across the transcript's
layout effects, plus the first-paint budget cut landing too late to
stop the full-budget commit.
Measured on the two largest local sessions (996 and 1363 messages),
main-thread longtask totals per switch: warm 2450ms -> 557ms and
1158ms -> 194ms; first paint 1690ms -> 444ms. Harness:
scripts/profile-session-switch.mjs (same CDP family as
profile-real-stream.mjs).
- use-resize-observer: drop the synchronous initial callback and ride
the observer's spec-guaranteed first delivery instead (same frame,
after layout, before paint). The sync call ran while the commit's
layout was dirty, so every size read in a callback forced a full
reflow — with one instance per user bubble (measureClamp read
scrollHeight, then WROTE --human-msg-full, re-dirtying layout for the
next bubble), the switch commit thrashed for over a second. Inside RO
timing the same reads are free. Composer metrics (2x
getBoundingClientRect + documentElement style writes) rides the same
fix.
- Same class, same fix at the remaining call sites profiling surfaced:
ExpandableBlock and TerminalOutput (dozens per tool-heavy transcript)
now measure/pin via RO initial delivery; the tool-window and
thinking-preview pins drop their sync pin() call; the thread
timeline's initial active-tick compute joins its existing
scroll-time rAF batching so back-to-back transcript updates coalesce.
- thread/list: cut the render budget in the RENDER phase (state-from-
props adjustment) instead of the post-commit layout effect. The
effect-time cut was too late — on a warm switch React first built and
committed the full 300-part tree, then re-rendered at 60, then bumped
back to 300, so the expensive commit still happened (and on a cold
switch the bump rAF usually fired while the transcript was still
empty, so the prefetched messages rendered at full budget anyway).
The render-phase cut restarts the component before any child renders;
a second trigger handles the cold path where messages land later
under the same sessionKey.
- thread/list: backfill 60 -> 300 inside startTransition so the older
turns' markdown+shiki render is interruptible background work instead
of a synchronous freeze one frame after the switch paints. Functional
Math.max so an urgent "Show earlier" click can't be rebased back down.
- composer focus: skip the rAF/timeout focus retries when the element
is already focused — focus() runs the full focusing steps (forcing
layout) even on the active element, ~585ms per switch on a large
dirty DOM.
- Replace the tautological render-budget test (it re-declared the
constants locally and asserted 60 < 300) with behavior tests of the
now-exported buildGroups + firstVisibleGroupIndex.
Verification: apps/desktop `npx tsc --noEmit` clean; full
`npx vitest run` 210 files / 1763 passed; manual CDP check confirms the
deferred backfill commits the full transcript, stays pinned to bottom,
and "Show earlier" still pages.
Post-hoc sprite removal is a losing battle (three inpainting strategies
failed QC on letter-edge overlap frames). The production answer: sprites
are BillboardComponent/SpriteComponent/ArrowComponent subobjects — set
bVisible:false via ObjectTools (remove_component fails on default
subobjects), swept scene-wide in one ProgrammaticToolset script
(148 actors, 13 sprites, one round-trip, verified).
New references/advanced-workflows.md covering the sophisticated-workflow
surface, each section exercised against a running editor:
- ProgrammaticToolset batching: full contract (get_execution_environment
gate, execute_tool fully-qualified names, JSON-string inputs,
returnValue unwrapping, allowed imports) + a verified worked example
(12-column colonnade, 36 components in one round-trip vs 37 serial calls)
- Blueprint DSL authoring loop, verified end-to-end: create -> list_graphs
-> get_graph_dsl_docs -> find_node_types per node -> write_graph_dsl ->
compile_blueprint -> spawn instance. Every node-ID gotcha hit live is
recorded (EventTick not Tick, Math|Rotator|MakeRotator, registry
categories vs doc categories, no (self) node)
- PIE session options schema (bSimulate/playMode/warmupSeconds/
startTransform, out-of-process downgrade behavior)
- Sequencer orientation: 140-tool surface mapped by capability group +
sibling keyframing/controlrig/conditions toolsets + minimal cinematic
skeleton
- LogsToolset self-debugging, AutomationTestToolset CI loop,
SemanticSearch, ConfigSettings, project AgentSkillToolset precedence
- Per-situation decision table
New pitfalls from this round's live failures: 10b (refPath-object vs
plain-string params; schema-in-error as tiebreaker), 10c (DSL node IDs
must come from find_node_types). SKILL.md: batching exception wired into
the operating loop, reference table row, description updated.
Ran the full loop against a real editor (blank project, ModelContextProtocol
+ ToolsetRegistry + AllToolsets enabled): raw MCP handshake, discovery walk,
environment relight for golden hour, primitive monument build, virtual-camera
captures with vision judgment, exposure debugging, annotated spatial capture.
67 toolsets advertised; every dispatch semantic below observed, not inferred.
Corrections and additions from the live run:
- Qualified toolset names (editor_toolset.toolsets.scene.SceneTools) with
SHORT tool_name; TOptional params must be explicit null; find_actors
requires ''/[] for its schema-required optionals; ObjectTools values is a
JSON *string*; refPath object references; returnValue wrapping; per-property
failure lists with schema-in-error
- HTTP wire contract: initialize=JSON + session header, tools/call=SSE frame
after game-thread completion (plain-JSON clients read empty body)
- CaptureViewport as virtual camera (captureTransform, meter-unit annotation
grid + actor callouts) verified with pixel evidence; recipes rewritten to
use it instead of viewport piloting
- New pitfalls from real failures: template-level environment-actor
duplication compounding into whiteouts (find-first/spawn-if-missing rule),
template exposure calibration vs physical lux (12b), objective exposure
check via ffprobe YAVG (12c), untitled-level Save-As modal deadlock,
macOS full-Xcode + Metal Toolchain requirement (xcodebuild
-downloadComponent MetalToolchain)
- Live toolset census (67 on blank project), LogsToolset/ConfigSettings/
SemanticSearch highlights, UE EULA 6(e) licensing note
Desktop launch and the update-chain rebuild install npm deps whose child
scripts shell out to a bare `node` (e.g. electron-winstaller's
select-7z-arch.js). When launched from the desktop updater chain
(Desktop -> hermes-setup -> hermes update) the shell PATH customizations are
lost, so the install dies with `'node' is not recognized` / `node: not found`.
- cmd_gui: wrap the npm-install env with with_hermes_node_path(_nixos_build_env())
so managed Node is prepended even on a stripped PATH — mirrors the idiom
already used by the update deps refresh. (The original fix merged nixos_env
on TOP of the managed env, whose full os.environ copy clobbered the managed
PATH back to bare; wrapping fixes that merge order.)
- _cmd_update_impl: spawn the `desktop --build-only` subprocess with
with_hermes_node_path() so the child starts with managed Node from the outset.
Regression test: the desktop install env now prepends the managed Node dir
ahead of a bare updater PATH instead of passing env=None.
Co-authored-by: F4TB0Yz <jfduarte09@gmail.com>
The web dashboard runs inside the gateway process, so `os.environ` carries
`_HERMES_GATEWAY=1`. `_spawn_hermes_action` spread that into the subprocess env,
so a spawned `hermes gateway restart` (dashboard "Enable webhooks", Telegram QR
apply) tripped the in-process restart-loop guard and exited 1 — the gateway
never restarted, but the dashboard reported `restart_started: true` because it
only checks that the spawn succeeded.
Scrub `_HERMES_GATEWAY` from the spawned action's env, matching what the
gateway's own restart watcher already does (gateway/run.py).
Fixes#52470. Adds a test asserting the spawned env drops the loop-guard var
while keeping HERMES_NONINTERACTIVE.
* fix(dashboard): unblock basic auth plugin during interactive password setup
When the dashboard prompts for username/password on a non-loopback bind,
also remove the bundled basic provider from plugins.disabled so
discover_plugins(force=True) can register it (#54489).
* test(dashboard): cover basic auth plugin blocked by plugins.disabled
Regression harness for #54489: credentials in config are not enough when
the bundled basic provider is on the deny-list.
Auto-compression ends the SessionDB session and forks a continuation,
rotating the stored session id. The gateway emits `session.info` with
the new `stored_session_id`, and the desktop's cache entry was updated
via `ensureSessionState` — but the URL route and `$selectedStoredSessionId`
never followed the rotation.
On the next send, `getRuntimeIdForStoredSession(oldStoredId)` returned
null (the cache entry's `storedSessionId` no longer matched the old id),
so `routedSessionNeedsResume` evaluated true, triggering a full
`session.resume` + REST transcript prefetch — the whole thread reloaded.
Fix: a new `$activeSessionStoredId` atom is set in `ensureSessionState`
when the active session's stored id changes. A `useEffect` in
`use-session-actions` subscribes to it and re-anchors the route +
selection (`setSelectedStoredSessionId` + `navigate(replace: true)`),
and cleans up the stale stored→runtime mapping.
`replace: true` because it's the same conversation — compression is
transparent to the user, so back-button stays correct.
* feat(desktop): add useKeybindHint hook and TipKeybindLabel
Add a shared hook that reads the current keybind combo for an action
id from the $bindings store (rebindable) or KEYBIND_READONLY (fixed),
returning a formatted string or null when unbound.
Add TipKeybindLabel — a convenience component that auto-reads both
its label (from i18n) and keybind combo from the action registry.
Pass only actionId for the common case; pass text to override when
the tooltip is context-dependent.
* fix(desktop): replace native title= on buttons with themed Tip
Migrate all <button>/<Button> elements using the native HTML title=
attribute to the instant, themed <Tip> component. Native tooltips are
unstyled, delayed (~500ms OS default), and visually inconsistent with
the app's instant themed tooltips.
Also adds <Tip> wrappers to icon-only buttons that were missing
tooltips entirely (dialog close, search clear, overlay close,
master-detail pane controls, keybind panel rebind/reset buttons).
Adds an enforcement test (no-native-title.test.ts) that scans all
.tsx files for <button>/<Button> with title= and fails if any are
found. Updates DESIGN.md with the icon-only button tooltip rule and
keybind hint guidance.
* feat(desktop): wire keybind hints into button tooltips
Add actionId to TitlebarTool, StatusbarItem, and SidebarNavItem so
their tooltips show the current keybind combo via TipKeybindLabel.
Fix the hardcoded NEW_SESSION_KBD in the sidebar to read from
$bindings so it stays live on rebind.
Wired surfaces:
- Titlebar: sidebar toggle, flip panes, keybinds, settings
- Statusbar: terminal toggle (view.showTerminal)
- Status stack: open agents button (nav.agents)
- Sidebar nav: new session, skills, messaging, artifacts
* feat(desktop): move keybind panel to settings tab with search filter
Move the keyboard shortcuts panel from a Radix Dialog into a proper
Settings tab (/settings?tab=keybinds). The ⌘/ shortcut and titlebar
keyboard button now navigate to this settings tab instead of toggling
a dialog. Adds a search filter to filter shortcuts by label.
- New: src/app/settings/keybind-settings.tsx (extracted from keybind-panel.tsx)
- Delete: src/app/shell/keybind-panel.tsx (dialog wrapper removed)
- Remove: $keybindPanelOpen atom and toggle/open/close functions
- Add: IconKeyboard to lib/icons.ts
- i18n: keybinds.search + settings.nav.keybinds (en, zh, zh-hant, ja)
* refactor(desktop): unify worktree dialog into shared WorktreeDialog
Extract the worktree creation dialog from StartWorkButton (sidebar) into a
shared WorktreeDialog component. Both the sidebar's StartWorkButton and the
composer's CodingStatusRow now use the same dialog, eliminating the duplicated
UI.
The shared dialog keeps the sidebar version's full feature set:
- BaseBranchPicker (filterable base branch combobox)
- Convert mode (check out an existing branch into a worktree)
- Sanitized branch name input
The coding row passes repoPath (from cwd) and onOpenWorktree (which carries
the composer draft to the new session) so the unified dialog works everywhere
there's a repo, not just inside an entered project.
Switching between chat sessions in the desktop app froze for up to ~1–2s on
large transcripts. Profiling the switch path surfaced three main-thread
blockers, fixed here minimally and without changing behavior.
1. JSON.stringify deep-compare (worst case). chatMessagesEquivalent compared
message parts with JSON.stringify(a) === JSON.stringify(b) on every switch.
On image-/large-blob-bearing transcripts this serialized every part twice
and cost well over a second. Replaced with a structural compare that never
stringifies: array-level identity fast-path, per-part reference fast-path,
then type-aware field comparison. The compare's only consumer asks "did the
transcript change, should I setMessages?", so it is deliberately
conservative — a false-negative just causes one extra idempotent
setMessages, while a false-positive (the unsafe direction) is avoided.
2. Scroll-settle loop. thread-list ran a requestAnimationFrame settle loop up
to 90 frames (or 5 stable frames) on every sessionKey change, each frame
forcing a synchronous layout read + write — racing the markdown paint for
up to ~1.5s. A normal synchronous switch stabilizes within a couple
frames, so the ceiling is now 2 stable frames / 15 max.
3. Synchronous first paint of up to 300 parts. On switch, thread-list reset
the render budget to the full RENDER_BUDGET=300, so up to 300 parts went
through markdown + shiki syntax-highlighting synchronously on the switch
commit. It now paints a small FIRST_PAINT_BUDGET=60 first, then bumps to
the full 300 in a requestAnimationFrame after the first commit.
Salvaged from PR #49807 by professorpalmer — re-applied to the restructured
file layout (use-session-actions/utils.ts, thread/list.tsx) and tests merged
into the existing utils.test.ts.
Co-authored-by: Cary Palmer <professorpalmer@users.noreply.github.com>
The sidebar working dot didn't update for background sessions until the
user opened them. Two coupled causes:
1. The gateway's session.info event payload omitted stored_session_id,
so the desktop app had no way to map a background session's runtime
id to its stored id. Without the stored id, setSessionWorking(null,
...) was a no-op — the $workingSessionIds atom never updated.
2. The running→busy transition in the session.info handler was gated on
`apply` (active session only). The gate correctly scopes view-only
side effects (setCurrentModel, setCurrentCwd, etc.) to the focused
chat, but the per-session busy state drives the sidebar indicator and
must reach every session. updateSessionState only mutates the
per-runtime cache entry, and syncSessionStateToView already guards
the view publish to the active session, so ungating is safe.
Fix: add stored_session_id to _session_info() in tui_gateway/server.py,
add the field to GatewayEventPayload, pass it to updateSessionState in
the session.info handler, and ungate the running→busy transition.
generate-patch now emits a has-fixes job output (true/false).
apply-patch is gated on it via `if: needs.generate-patch.outputs.has-fixes == 'true'`,
so when `npm run fix` produces no diff, the privileged job is skipped
entirely — no runner allocation, no redundant checkout/download/push/PR.
Replaces moonshotai/kimi-k2.6 (recommended) and moonshotai/kimi-k2.7-code
with moonshotai/kimi-k3 in both curated lists, regenerates the published
model-catalog.json manifest, and updates the docs example manifest (en+zh).
kimi-k3 verified live on both endpoints (Nous Portal /v1/models and
OpenRouter /api/v1/models; 1M context, $3/$15 per Mtok). Family-prefix
matching already covers k3 in moonshot_schema, cache policy, and context
heuristics — no code changes needed there.
queryRewrite (default off) gates the latest-message rewrite so the
extra auxiliary LLM call is opt-in. firstTurnBaseWait and
firstTurnDialecticWait expose the turn-1 bounded waits in seconds
(0 disables). All three resolve host-block-first like every other
field. Also pins per-host timeout resolution with tests.
Move query_rewrite from the honcho plugin to plugins/memory/ and
rename the auxiliary task key honcho_query_rewrite ->
memory_query_rewrite so any memory provider can use the same
rewrite path and model/timeout config block. No behavior change.
dialecticMaxChars (default 600) is documented as the budget for the dialectic
supplement auto-injected into the system prompt every turn — a small recurring
cost that is correct to bound tightly. But dialectic_query() applied that cap
unconditionally, so explicit honcho_reasoning tool calls — where the model
deliberately spends a turn asking for a synthesized answer — were silently
truncated mid-word to 600 chars with a trailing " …", no error surfaced. The
full answer is returned by Honcho server-side; the clip happens client-side.
The auto-injection path already has its own token-based budget (contextTokens,
enforced in prefetch() via _truncate_to_budget), so the char cap's real job is a
cheap always-on guardrail for that recurring injection. Explicit tool results
are already bounded server-side by Honcho's dialectic MAX_OUTPUT_TOKENS and don't
need the injection cap — sibling tools (honcho_search, honcho_context) don't
post-clip their results either.
Add apply_injection_cap (default True, preserving current behavior) to
dialectic_query(); the honcho_reasoning tool handler passes False so it returns
Honcho's full synthesized answer. Auto-injection is unchanged. Tests cover both
the capped injection path and the uncapped tool path.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
honcho_conclude's delete action was unreachable in practice: no tool ever
surfaced a real conclusion id for the model to pass as delete_id.
honcho_search only searches the separate Message resource space, and the
SDK's ConclusionScope.list()/.query() (which do return real Conclusion.id
values) were never wired into any tool.
Adds an optional list mode to honcho_conclude (query to search, omit to
browse recent conclusions), backed by a new
HonchoSessionManager.list_conclusions(). No new tool, no changes to the
create/delete signatures or their conclusions_of() routing.
injectionFrequency='first-turn' returned empty for the entire
prefetch_context() method on turns 2+, which blocked the dialectic
supplement from being consumed and injected. The dialectic has its
own cadence (dialecticCadence) and must continue to fire and inject
independently of the base context layer.
Now first-turn mode gates only Layer 1 (base context: representation
+ card), letting Layer 2 (dialectic supplement) flow through its
normal consumption path on every turn.
Also fixes all remaining tests that passed dialecticCadence via
cfg_extra={'raw': {...}} to use the typed dialectic_cadence field.
injectionFrequency, contextCadence, and dialecticCadence were read
only via raw.get() which checks root-level keys in honcho.json.
Settings placed inside hosts.<name> (the normal per-host location)
were silently ignored, falling back to defaults.
- Add typed dataclass fields: injection_frequency, context_cadence,
dialectic_cadence with host-block-first resolution chains matching
the pattern used by all other config fields.
- Update HonchoMemoryProvider.initialize() to read from typed cfg
fields instead of cfg.raw directly.
- Fix search_context tests that mocked _honcho directly — the
.honcho property getter calls get_honcho_client() which overwrites
the backing field, so use patch.object on the property instead.
- Add injection_frequency and context_cadence config override tests.
HonchoClientConfig timeout/requestTimeout resolution skipped the per-host config
block, silently dropping a host-scoped timeout and falling through to the global
config.yaml value (or the default). Add the host block at the front of the
resolution chain, consistent with every other field (base_url, api_key, etc.).
Symptom: Honcho logs show a dialectic answer was generated, but Hermes never
injects it — intermittently.
Root cause: the dialectic supplement that queue_prefetch() fires at the end of
turn N is stored pending (fired_at=N) for consumption by turn N+1's prefetch().
But prefetch()'s trivial-prompt guard returned early BEFORE the consumption
block. So when turn N+1's prompt was trivial ('ok', 'yes', 'continue', a slash
command), the ready result was never consumed, and a few turns later the
stale-discard guard dropped it. Generated by Honcho, never seen by the model.
The dependence on 'is the consuming turn trivial?' is why the loss looked random.
Fix: trivial turns now consume and inject a ready, non-stale pending result while
still spending no new work (no base-context fetch, no new dialectic fire). A
trivial ack shouldn't generate context, but it shouldn't destroy an answer
already computed for that exact turn. Extracted the pop+stale-check into a shared
_consume_pending_dialectic() used by both the trivial path and the normal path so
they age-check identically.
Preserved (covered by new tests): genuinely stale results are still discarded on
trivial turns; trivial turns still fire no new work; a trivial turn with nothing
pending still injects nothing.
Adds regression tests for inject-on-trivial and discard-stale-on-trivial.
A brand-new session injected no Honcho context on the user's first message —
the peer card/representation only showed up from turn 2 onward. The base-context
fetch was fired asynchronously and popped in the same synchronous pass, so it
always lost the race on turn 1 (a background thread can't finish inside one
pass), leaving the first response with zero recalled context.
Fetch the base layer (representation + card + summary) synchronously with a
bounded timeout on turn 1 so the peer card is injected immediately; subsequent
turns still consume the background-refreshed result primed by queue_prefetch().
The wait is bounded by _FIRST_TURN_BASE_TIMEOUT and tightened further by a small
configured request timeout (fail-fast deployments / tests).
Two related first-turn/dialectic reliability fixes ride along:
- first-turn dialectic no longer double-fires: if a prewarm .chat() thread is
already in flight from session init, turn 1 waits briefly for it instead of
firing a second (duplicate) call that also blocked the first response. The
first-turn wait is decoupled from a large host timeout (a 60s host timeout
must not block the first response for 60s) via _FIRST_TURN_DIALECTIC_CAP,
while still honoring a tight configured timeout.
- empty-pass propagation guard in multi-pass dialectic: at depth > 1 each pass
feeds the prior pass's output into the next prompt. If a pass returned empty
(e.g. a reasoning model that spent its whole budget thinking), the next prompt
carried a blank assessment (the "empty spot" seen in Honcho request logs). Now
only non-empty prior results feed dependent passes; if all priors are empty,
re-issue the base prompt instead of referencing nothing.
The five Honcho tool schemas had overlapping/misleading descriptions, making it
hard for the model to pick the right one, plus two concrete correctness bugs:
- honcho_context advertised an 'Optional focus query' parameter that the dispatch
never read. The model could pass query= expecting filtering that never happened.
Remove the dead parameter; honcho_context is an honest no-query snapshot. Focused
retrieval now lives in honcho_search (see prior commit).
- honcho_conclude's peer param was described as 'Peer to query' — wrong; it's the
peer the conclusion is ABOUT. Corrected.
Rewrite all five descriptions to give each tool a distinct mental model and
cross-reference siblings:
profile = read/write the compact card (cheapest, no query, no LLM)
search = find what was actually said, ranked, cross-session (cheap, no LLM)
reasoning= ask a question, get a synthesized answer (the only LLM tool; expensive)
context = a fixed session snapshot (no query, no LLM)
conclude = write a durable fact to the profile
Addresses the honcho_context query param half of #29402 (see PR notes for the
divergence from that issue's proposed wire-through approach).
honcho_search routed through search_context() -> peer.context(search_query=),
which returns the peer's standing representation + card. The search_query arg
does not turn that endpoint into a search, so results were effectively
query-independent: the same representation blob regardless of the query. Factual
lookups ('what medication', 'which value did we pick') returned noise.
Rewire search_context() to call the workspace message-search endpoint
(Honcho.search) with a peer_perspective filter: RRF-ranked (hybrid semantic +
full-text) raw message excerpts spanning every session the peer was a member of,
across all authors, membership-time-scoped. This is the cross-session factual
recall primitive.
peer_perspective is chosen over the alternatives because it is the only scope
that is simultaneously (a) cross-session, (b) inclusive of assistant-authored
facts about the peer (peer-author search drops these, and they are a large part
of what you want to recall about yourself), and (c) privacy-scoped to the peer's
own sessions (plain workspace search leaks other peers' sessions).
- snippets are labeled by author so the model can tell user-stated facts from
assistant-derived ones
- max_tokens is now an enforced budget (was accepted but meaningless)
- graceful fallback to peer-authored search if peer_perspective is unsupported
- query length clamped under the embedding input cap
Replaces 3 change-detector tests that asserted the old representation-dump
behavior with 4 that assert the message-search contract + fallback path.
auto_title_session runs as a bare daemon-thread target. Any exception
escaping it hits the default threading excepthook and sprays a raw
traceback into the user's terminal mid-session. The canonical trigger
is the post-'hermes update' stale-module window: the function's lazy
imports read NEW source from disk while already-imported modules
(agent.portal_tags) are still the OLD cached version, producing an
ImportError that repeats on every auto-title attempt until the
long-running process restarts (seen live after 9ce0e67f2 added
set_conversation_context).
The public entrypoint now wraps the body in a catch-all that logs one
WARNING naming the likely cause ('restart the running Hermes process'),
routes the exception through the existing failure_callback channel
(user-visible warning in CLI, debug-suppressed in gateway per #23246),
and never re-raises. This also makes the function honor its own
docstring contract ('silently skips if title generation fails').
react-router v7's HashRouter wraps every route state update in
React.startTransition() by default. In React 19's concurrent renderer,
transitions are non-urgent — React can yield mid-render and come back
later. When the app is under load (streaming token deltas, gateway
events, store updates from active sessions), those higher-priority
updates keep interrupting the transition, starving the route change commit.
This matches every symptom of the session-switch lag:
- Main thread is free (animations run, clicks work) — startTransition
defers, it does not block
- navigate() does not take effect for seconds — the transition keeps
getting interrupted by higher-priority updates
- Worse under load — more concurrent re-renders = more interruptions
- The whole UI (sidebar + main pane) does not update — the entire route
change is one transition, so nothing commits until it finishes
Pass useTransitions={false} to HashRouter so route state updates are
synchronous at default React priority instead of deferred transition
priority. navigate() now commits and paints immediately.
See: react-router v7 chunk-BIP66BKV.js HashRouter implementation.
Add the flag-gated compute-host supervisor, delta/control protocol, PPID orphan guard, inline fail-open path, synthetic GIL-heavy turn seam, and AC-4 certify harness.
Verified on current origin/main: 343 focused tests pass; Ruff and diff checks pass; 360s AC-4 run with six heavy lanes passes at 6.11ms serving p99 with zero stalls and valid load.
Co-authored-by: Kyzcreig <9063726+Kyzcreig@users.noreply.github.com>
Reassert the persisted webContents zoom after BrowserWindow moved, covering Windows monitor transitions where Chromium recalculates display scaling and drops the user-selected zoom.
Rebased onto current main, where the electron test harness migrated from
node:test to vitest (test:desktop:platforms = `vitest run --project electron`,
auto-discovering electron/*.test.ts). Swap the node:test import for vitest and
drop the .ts-extension import hack; the obsolete package.json node --test list
edit is dropped in the cherry-pick resolution since vitest auto-discovers.
Co-authored-by: Gille <4317663+helix4u@users.noreply.github.com>
Bare /reasoning and /fast now render a native one-tap picker on
picker-capable platforms, with automatic fallback to the text status
card everywhere else — parity with the /model picker UX.
- gateway/slash_commands.py: generic send_choice_picker capability gate
(detected on the adapter type, like send_model_picker); selection and
typed arguments flow through one shared application path so they can
never diverge; choices built from VALID_REASONING_EFFORTS so future
levels appear automatically
- telegram: flat inline-keyboard picker (cp:<idx> callbacks), authorized
users only (same gate as approval buttons)
- discord: ChoicePickerView select menu, auth mirrors ExecApprovalView,
2-minute timeout with expiry edit
- matrix: reaction-based picker; reaction set extended to 12 slots to
fit the full effort ladder + subcommands
- locales: picker_title + choice labels in all 16 languages
- docs: ADDING_A_PLATFORM.md capability table
Closes#61110.
selectModel in use-model-controls captured activeSessionId as a closure
prop, but the actions bag in wiring.tsx mutates in place to keep a stable
identity for memoized surfaces. The modelMenuContent useMemo captures
selectModel once when the gateway first opens (before any session is
active) and never re-evaluates, so clicking a model in an existing thread
goes through a stale closure with activeSessionId=null — the pick is
treated as UI-only, config.set is never sent, and the next session.info
event clobbers the optimistic update back to the session's real model.
Drop the activeSessionId prop entirely. All three callbacks now read
$activeSessionId.get() live from the store, matching the pattern
refreshCurrentModel already followed. This is the correct contract for
the actions bag's in-place mutation: callbacks read live state from the
store, not from captured props.
The salvaged choices list predates the max/ultra effort levels (#62650);
add them so the Discord dropdown matches the canonical ladder. Discord
caps choices at 25 — we're at 11, plenty of headroom.
The Discord /reasoning command declared a single free-text `effort`
parameter, so the native UI funneled every invocation into that one box
and never surfaced the reset / show / hide subcommands the gateway
handler already supports. Replace the free-text param with an explicit
choices dropdown covering the effort levels plus reset/show/hide,
mirroring the existing /tokens and /voice commands. --global persistence
stays reachable by typing the command as plain text.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
terminal_tool reads all settings from TERMINAL_* env vars, bridged from
config.yaml by the CLI, gateway, and TUI-PTY launchers. Processes that
skip every launcher bridge — hermes serve / the Desktop app backend's
in-process agents, the desktop cron ticker — saw an unset TERMINAL_ENV
and silently ran every command on the host even when config.yaml selects
terminal.backend: docker. A user who configured Docker isolation got
unsandboxed host execution with no warning.
Two layers:
- _ensure_terminal_env_bridged() in _get_env_config(): when TERMINAL_ENV
is unset, backfill TERMINAL_* from config.yaml via
apply_terminal_config_to_env(override=False). Explicit env always wins
(honor explicit choice; only fix the accidental fallback). One-shot,
fail-open to the historical local default.
- cmd_dashboard/serve: run the same bridge at startup so every consumer
in the backend process (in-process agents, desktop cron ticker,
tui_gateway cwd resolution) sees the bridged env directly.
Fixes#63141, #54449, #61115, #65696.
hermes update ran TWO separate pre-update backup mechanisms: the
config-gated full zip (updates.pre_update_backup, default off) and an
unconditional quick state snapshot added for #15733 that ignored the
user's setting entirely. On a large state.db (observed: 24 GB) the
'cheap' snapshot silently added ~60s to every update and ate 24 GB of
disk in state-snapshots/.
Now there is ONE mechanism, gated by updates.pre_update_backup with
three modes:
- quick (new default): state snapshot of critical small files (pairing
JSONs, cron jobs, config, auth, per-profile DBs). Files over 1 GiB
are skipped with a warning so a bloated state.db can never stall the
update again.
- full: the quick snapshot plus the HERMES_HOME zip (old 'true'
behavior; --backup forces it for one run).
- off: nothing runs — an explicit opt-out now disables the quick
snapshot too (--no-backup does the same per-run).
Legacy booleans are honored: true -> full, false -> off.
_run_pre_update_backup() now returns the quick-snapshot id so the
post-update cron-jobs restore safety net (#34600) keeps working; the
snapshot moved from the post-fetch site to the pre-mutation site,
which also covers the zip-fallback update path it previously missed.
Add current-main regression coverage for both the registered provider
profile and core GitHub Responses path while leaving live catalog loading
to the complementary catalog-resolution work in #51953.
The Copilot provider profile unconditionally mapped ``xhigh`` to ``high`` before
checking the model's catalog, so models that DO support ``xhigh`` (e.g. the
gpt-5.x family per the live /models catalog) were silently capped one level
down.
Honor the requested effort when the catalog lists it as supported, and only
downgrade when it does not, choosing the nearest weaker supported level
(xhigh->high, minimal->low, else medium, else the first supported level). This
matches the nearest-down clamp behavior used elsewhere for the ``max`` effort.
Adds tests/plugins/model_providers/test_copilot_profile.py covering forward,
downgrade, and fallback paths (catalog lookup stubbed).
The current code unconditionally downgrades 'xhigh' to 'high' whenever
'high' is in supported_efforts, even if 'xhigh' is also supported.
This prevents users from using extended thinking on providers like
Copilot that list 'xhigh' in their supported efforts.
Fix: only downgrade 'xhigh' to 'high' when 'xhigh' is NOT in the
provider's supported efforts list.
The desktop self-update rebuild (`hermes desktop --build-only`, driven by
the macOS in-app updater) runs `_run_npm_install_deterministic`, which used
plain `npm ci` / `npm install`. Those honor an inherited
`NODE_ENV=production` (or npm `omit=dev`) and silently omit devDependencies.
The desktop build toolchain — tsc, vite, electron-builder — are all
devDependencies, so under `NODE_ENV=production` the install completes (exit 0)
but `tsc` is never placed, and the very next step (`tsc -b && vite build`)
dies with `tsc: command not found` (exit 127). The user sees
"UPDATE DIDN'T FINISH / Rebuilding the desktop app failed (exit 127)" even
though the git/pip update applied cleanly. NODE_ENV=production can leak in
from a shell profile, a parent process, or a packaged-app launch context.
Force `--include=dev` on both the `npm ci` and `npm install` paths. The only
callers are frontend builds (desktop / TUI / web), which always need the dev
toolchain, so this is safe across the board.
Verified empirically: under NODE_ENV=production, `npm ci` leaves tsc MISSING
while `npm ci --include=dev` installs it. Added two regression tests
(test_npm_ci_forces_include_dev, test_npm_install_fallback_forces_include_dev)
— both confirmed to FAIL when the fix is reverted.
When no agent is active, '/queue <prompt>' previously fell through
dispatch with its raw text intact instead of being treated as a
normal prompt. Rewrite event.text to the bare payload (mirroring the
/steer no-active-agent path just below) and return a usage hint when
the payload is empty.
Salvaged from PR #29290 (queue half only — the /footer mid-run
dispatch half already landed on main via #65521).
Three follow-up fixes to the salvaged reasoning_effort support, all verified
live against ollama.com /v1/chat/completions + /api/show on deepseek-v4-pro,
gemma3, and qwen3-coder:
1. Capability-gate on /api/show 'thinking'. The original ignored the
supports_reasoning flag and emitted reasoning_effort for every model. Now
gated: only models whose native /api/show capabilities list contains
'thinking' (deepseek-v4 yes; gemma3 / qwen3-coder no) get reasoning_effort.
Mirrors the LM Studio pattern — capability resolved once per (model,
base_url) in run_agent._supports_reasoning_extra_body via a cached probe
(hermes_cli.models.ollama_model_supports_thinking), threaded into the
profile hook as supports_reasoning. No live HTTP in the per-request path.
2. Disable actually disables. Ollama Cloud defaults to thinking ON and IGNORES
the extra_body.thinking:{type:disabled} shape (verified: still returned
reasoning). The only working off switch is top-level reasoning_effort:'none'.
The salvaged code returned ({}, {}) for enabled:false / effort:none, leaving
thinking ON. Now emits {'reasoning_effort': 'none'}.
3. Omit unrecognized effort. The original forwarded any unknown string verbatim
including 'minimal' (a real Hermes effort level). Ollama Cloud rejects
unrecognized values with a hard HTTP 400 (accepted set: low/medium/high/
max/none), so forwarding 'minimal' would break the request. Now omitted.
Core touches (run_agent.py, hermes_cli/models.py) add the capability probe;
the plugin profile only consumes the resolved flag. 24/24 profile tests green;
194 provider/transport tests unaffected.
LM Studio's request vocabulary tops out at "xhigh", but Hermes' generic
effort ladder has since grown two stronger levels. "max" and "ultra" miss
the _LM_VALID_EFFORTS membership test, keep the initialized "medium"
default, and are thereby conflated with unparseable input -- so asking for
more reasoning yields less than "xhigh":
high -> 'high' xhigh -> 'xhigh'
max -> 'medium' ultra -> 'medium'
This is drift, not a design choice. The valid set was an exact mirror of
VALID_REASONING_EFFORTS when the file was authored; the ladder then grew
"max" and later "ultra", and the sweep that taught every other provider
about the new levels missed this module -- it has never been touched since
it was written.
Clamp the two stronger levels onto LM Studio's declared ceiling instead,
mirroring the ceiling clamp every other provider already applies. Widening
_LM_VALID_EFFORTS would instead assert that LM Studio accepts "max" on the
wire, which is a provider-side claim this repo cannot verify; clamping
consumes only the ceiling the file already declares for itself.
The clamp is kept separate from _LM_EFFORT_ALIASES because that mapping is
also applied to the model's published allowed_options, which must not be
rewritten. A clamped value stays subject to the allowed_options check, so a
model that does not publish "xhigh" still gets the field omitted and falls
back to its own default -- exactly how a directly-requested "xhigh" behaves.
The regression test asserts monotonicity over the canonical ladder rather
than the two values alone, so the next level added upstream cannot silently
reintroduce the inversion.
Replaces the salvaged tests' dict-config path (which required a
GatewayRunner.__init__ dict-coercion hack — dropped from this salvage;
the second commit on the PR branch existed only to support it) with the
established bare-runner test pattern. Also asserts full argument
passthrough to the inner task.
When multiplex_profiles is true, background tasks spawned by /background
command failed with UnscopedSecretError because _resolve_session_agent_runtime()
was called without a profile secret scope. This fix wraps the task in
_profile_runtime_scope, mirroring the pattern used by _run_agent.
Fixes#60726
PR #36019 documented the attachment tools in the kanban-worker skill,
but main removed that skill in #50473 and folded its content into the
KANBAN_GUIDANCE prompt block. Land the same guidance there instead so
every dispatcher-spawned worker sees it.
_download_url_with_cap called urlopen() after only a scheme check, so a
model-controlled URL could reach loopback services, RFC1918/CGNAT hosts,
or cloud metadata endpoints (169.254.169.254), and a public host could
302 to any of those unvalidated.
Route the fetch through the repo's canonical SSRF guard instead:
validate every hop with tools.url_safety.is_safe_url() and follow
redirects manually (httpx, follow_redirects=False, 5-hop limit) so each
Location target is re-checked before it is fetched — the same pattern
as tools/skills_hub._guarded_http_get. The streaming size cap is
unchanged. Local-fixture tests opt in via HERMES_ALLOW_PRIVATE_URLS
(the guard's documented escape hatch); new tests pin rejection of
loopback, cloud-metadata, and private-range URLs, a mocked
public→loopback redirect, and a mocked public happy path.
The salvaged attachment-toolset commit predated main centralizing the
25 MB cap as kanban_db.KANBAN_ATTACHMENT_MAX_BYTES and re-introduced a
private _MAX_ATTACHMENT_BYTES alias. Drop the duplicate: kanban_db's
store_attachment_bytes(), the dashboard upload endpoint, and the
kanban_attach_url tool all reference the one shared constant now, and
the tests monkeypatch that same name.
The kanban board has had full attachment storage and a dashboard HTTP
API (upload/list/download/delete) since #35338, but there was no agent
toolset tool and no `hermes kanban` CLI verb for attachments. Agents and
scripts that don't go through the dashboard server (or can't touch the DB
directly) had no way to create or read real attachments — only links in
comments.
Close that gap by mirroring the existing comment surface:
- `kanban_db.store_attachment_bytes()` — one shared write path (validate
name, enforce the 25 MB cap, write the blob under the per-task dir with
collision-free naming, insert the metadata row, clean up an orphan blob
if the insert fails). `_MAX_ATTACHMENT_BYTES`, `_safe_attachment_name`,
and a new `_collision_free_path` move here so the dashboard, the tool,
and the CLI all share one implementation and can't drift.
- Tools (`tools/kanban_tools.py`): `kanban_attach` (inline base64),
`kanban_attach_url` (server-side http/https fetch with the same cap),
`kanban_attachments` (list). Write tools respect worker task-ownership;
list is read-only. Registered in the `kanban` toolset.
- CLI (`hermes_cli/kanban.py`): `attach <id> <path>`, `attachments <id>`,
`attach-rm <attachment_id>`.
- Dashboard `upload_task_attachment` now imports the shared helpers and
uses `_collision_free_path` — behavior identical (still streams to disk
with the cap, still 413 on overflow).
- Docs (AGENTS.md, kanban-worker skill) and toolset membership updated.
Tests: tool round-trip + oversize + bad base64 + ownership; attach_url
against a local HTTP fixture incl. oversize-mid-stream and non-http
scheme rejection; CLI attach/attachments/attach-rm; shared-helper unit
tests; dashboard parity preserved.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The platform callback verifier can do blocking network I/O (e.g. the
google-chat adapter fetches Google signing certs on a cache miss), which
would stall the event loop if called inline. Run sync verifiers via
asyncio.to_thread (await coroutine verifiers directly), and treat a
crashing verifier as a 401 rather than a 500 through the dispatch path —
a broken verifier must never admit an event.
DeepSeek's own API (api.deepseek.com) reports context-cache hits as
top-level usage.prompt_cache_hit_tokens / prompt_cache_miss_tokens
(prompt_tokens = hit + miss), not the OpenAI nested
prompt_tokens_details.cached_tokens shape. Neither normalize_usage()
nor the chat_completions transport's extract_cache_stats() read those
fields, so direct DeepSeek sessions always showed 0 cache-hit tokens:
invisible in accounting, mis-billed at the full input rate, and 0%
cache display.
Both layers now fall back to prompt_cache_hit_tokens when the nested
shape is absent; the nested value wins when both are present (proxies).
Fixes#61871.
Server-side coverage for the critical-CSS shim (PR #36024 salvage):
- user theme → style block emitted with ONLY real bundle variable names
(--background-base/--midground-base from layerVars(),
--theme-font-sans/--theme-base-size from typographyVars()/index.css),
and an html,body rule expressed via those vars so runtime theme
switches never leave a stale canvas/font
- built-in / unknown / non-string active theme → no block
- malformed theme YAML and load_config() exceptions → no crash, index
still serves
- </style> breakout attempt in a theme value stays escaped
- mount_spa integration: block present in <head> for user themes,
absent for built-ins
Review fixes for the inline critical-CSS bootstrap (PR #36024):
1. Variable names now match what the bundle actually consumes.
--color-background, --color-midground, --font-sans and
--font-base-size appear nowhere in web/src; the real tokens are:
--background-base / --midground-base (layerVars(), context.tsx)
--theme-font-sans / --theme-base-size (typographyVars(), and
index.css html{font-family:var(--theme-font-sans);
font-size:var(--theme-base-size)})
2. Stale-rule bug: the injected html,body rule previously baked in
literal hex/font values. Because the <style> block sits after the
bundle's <link> at equal specificity and is never removed, switching
themes in the picker left the old canvas/font until reload. The rule
now references the same CSS variables instead of literals —
applyTheme() writes those vars as inline styles on documentElement,
which outrank this block in the cascade, so runtime theme switches
re-resolve the rule automatically. No frontend change needed.
User themes (`~/.hermes/dashboard-themes/*.yaml`) reach the SPA only
after `/api/dashboard/themes` resolves at React mount. The bundle paints
the first frame with the default Hermes Teal canvas — the
`<link rel="stylesheet">` carries `:root{--background-base:#041c1c}`,
the bundled `presets.ts` defines the same surfaces in JS — and then
`ThemeProvider.applyTheme(<user theme>)` flips the inline CSS variables
on `documentElement` once the API response lands. Visible to the user
as a green canvas behind the loading SPA on every reload when the active
theme is non-default.
Built-in themes do not suffer the same effect because their full
definitions ship inside the bundle, so the SPA already has the palette
before first paint.
This patch closes the gap on the backend side: `_serve_index()` injects
a `<style id="hermes-theme-bootstrap">` block inside `<head>` with the
six critical CSS variables (`--background-base`, `--color-background`,
`--midground-base`, `--color-midground`, `--font-sans`,
`--font-base-size`) plus an `html, body` rule painting the body in the
target palette. Because the inline `<style>` follows the bundle's
`<link>` in DOM order and matches the same `:root` specificity, the
later declaration wins the cascade — the static canvas behind the SPA is
already the right colour before any JavaScript runs.
`_render_active_theme_bootstrap_css()` looks up the active theme through
the existing `_discover_user_themes()` helper. No-op for built-in
active themes (empty string returned, no `<style>` injected). No new
API endpoints, no config flags, no frontend changes.
After `ThemeProvider` mounts and `applyTheme()` writes the same
variables as inline styles on `documentElement`, the values match what
the bootstrap block set, so there is no second-paint discrepancy on the
critical CSS variables.
Review follow-ups on the cherry-picked #36043 commit:
1. Guard the custom:<name> passthrough with a _get_named_custom_provider
lookup. The PR unconditionally kept the full custom:<name> string, which
broke config-less runtime custom providers (#34777 regression — entries
that exist only in the live runtime, not config.yaml): the named arm
found no entry and resolution fell through to Step 2. Now custom:<name>
only takes the named arm when a config entry actually exists; otherwise
it collapses to the anonymous-custom arm with the runtime endpoint,
preserving pre-PR behavior.
2. Drop the dead 'explicit_api_key = runtime_api_key' assignment (and its
misleading comment) in the named-entry branch. resolve_provider_client's
named-custom arm derives the key exclusively from the entry's
api_key/key_env and never reads explicit_api_key, so the assignment was
a no-op. Wiring precedence in was not justified: for a named custom
provider the runtime key IS the entry's key (set_runtime_main sources it
from the same config), so deletion is the honest option.
3. Tighten the Palantir Bearer-auth check from a loose substring match
('palantirfoundry' in normalized) to a hostname match via
base_url_host_matches(..., 'palantirfoundry.com'), so path segments or
lookalike domains containing the string no longer trigger Bearer auth.
Tests: named-custom anthropic_messages end-to-end routing (full name kept,
AnthropicAuxiliaryClient at the original /anthropic URL, no /v1 rewrite)
plus Palantir Bearer-auth positive and substring-false-positive cases.
When the user's main provider is a named custom_providers entry exposing an
Anthropic Messages surface (e.g. Palantir Foundry's
/api/v2/llm/proxy/anthropic, custom LiteLLM/Bedrock proxies), auxiliary
tasks (title generation, compression, web extract, session search, etc.)
returned HTTP 404 NOT_FOUND for every call.
Root cause: `_resolve_auto` collapsed any `custom:<name>` main provider
to plain `"custom"` and passed runtime_base_url as explicit_base_url.
This landed in `resolve_provider_client`'s anonymous-custom arm
(`if provider == "custom":`), which unconditionally calls
`_to_openai_base_url` — that helper strips a trailing `/anthropic` and
substitutes `/v1` (designed for MiniMax/ZAI which expose both surfaces).
The result for Palantir is `/api/v2/llm/proxy/v1`, which does not exist
on the proxy — every auxiliary call 404s. The runtime `api_mode=
anthropic_messages` flag was discarded by this arm.
Fix: split the conditional so only the literal `"custom"` provider takes
the anonymous-custom path; `custom:<name>` keeps its full `custom:<name>`
string when handed to `resolve_provider_client`, where the
named-custom-provider arm (added in earlier work) honours the entry's
`api_mode` and routes through `AnthropicAuxiliaryClient` against the
original `/anthropic` URL.
Also: extend `_requires_bearer_auth` in `anthropic_adapter.py` to
recognise palantirfoundry hosts so the SDK sends `Authorization: Bearer`
instead of the default `x-api-key` (Palantir's proxy rejects x-api-key
with 401).
Verified end-to-end against a live Palantir Foundry deployment with both
claude-4-6-opus and claude-4-7-opus models — `generate_title` returns
real titles instead of 404ing. Regression-tested:
- anonymous `custom` (with base_url) still routes to OpenAI wire
- built-in NVIDIA provider unchanged
- custom-without-base_url still falls through to Step-2 chain
Same bug class as the salvaged #65305/#65307: hmac.compare_digest (and
secrets.compare_digest) raise TypeError when given a str containing
non-ASCII characters, and these call sites feed it raw request input.
Compare as UTF-8 bytes everywhere:
- gateway/platforms/msgraph_webhook.py: clientState from request body
- gateway/platforms/whatsapp_cloud.py: hub.verify_token query param +
X-Hub-Signature-256 header (comment claimed 'works on str' — it
doesn't for non-ASCII)
- plugins/platforms/feishu: verification token + x-lark-signature
- plugins/platforms/raft: bridge token header
- plugins/platforms/line: X-Line-Signature
- plugins/platforms/sms: X-Twilio-Signature
- tools/code_execution_tool.py: sandbox RPC token (both loops)
Regression tests for the two gateway-core sites (msgraph, whatsapp).
The fix routes the Svix v1 comparison through _hmac_str_equal too, but the
existing non-ASCII tests only exercised the GitHub/GitLab/generic V1/V2
branches. Add a Svix case (valid svix-id + fresh svix-timestamp so it
reaches the v1,<sig> compare) with a non-ASCII signature, which raised
TypeError before the fix and now rejects cleanly.
_validate_signature backs the public webhook receiver. It compared each
attacker-supplied signature/token header (GitHub X-Hub-Signature-256,
GitLab X-Gitlab-Token, generic X-Webhook-Signature / -V2, and the Svix v1
header) against a computed hex/base64 digest with hmac.compare_digest on
two str values. compare_digest raises TypeError on a str containing
non-ASCII characters, and the header is raw client input on an
unauthenticated endpoint — so any internet client could POST a single
non-ASCII byte in the signature header and raise out of the handler,
returning a 500 instead of a clean 401. Fail-closed, but an on-demand
crash of the request path.
Route all five comparisons through a small _hmac_str_equal() helper that
encodes both sides to UTF-8 bytes before the constant-time compare
(compare_digest has no ASCII restriction on bytes). Semantics are
unchanged for valid signatures; a hostile non-ASCII header now fails
closed with a rejection instead of raising.
Adds regression tests: non-ASCII GitHub/GitLab/generic/V2 signature
headers return False (no raise), and a non-ASCII configured secret still
matches its exact token value.
Also maps drexux0@gmail.com in scripts/release.py AUTHOR_MAP.
_check_auth gates every OpenAI-compatible API server endpoint. It compared
the client's raw bearer token against the configured key with
hmac.compare_digest on two str values. compare_digest raises TypeError on
a str containing non-ASCII characters, and the token comes straight from
the Authorization header — so a request with a single non-ASCII byte in
the key (a stray unicode char, a smart quote, a pasted BOM) crashed the
handler with an unhandled TypeError. Every endpoint calls _check_auth
without a try/except, so the framework turned that into a 500 Internal
Server Error instead of the intended 401 Invalid API key.
Compare as bytes, matching web_server.py's dashboard-token check
(hmac.compare_digest(auth.encode(), expected.encode())). Encoding both
sides keeps the timing-safe comparison and its semantics identical for
valid keys while making a non-ASCII token fail closed with a clean 401.
Adds regression tests: a non-ASCII bearer token returns 401 (no raise),
and a non-ASCII configured key still authenticates against its exact
value.
Follow-up to the truncate_message split-loop floor. Two review points:
- CapabilityDescriptor.from_json trusted the wire max_message_length
verbatim, so a connector advertising 0 ('no limit') — or a buggy one
sending 0/negative — produced a descriptor whose bound flowed straight
into the adapter's MAX_MESSAGE_LENGTH and truncate_message. Normalize
it to the documented 4096 default (mirrors from_platform_entry's
'or 4096' and docs/relay-connector-contract.md), fixing the degenerate
budget at its source rather than only surviving it downstream.
- Document the truncate_message length contract for a budget too small
for one codepoint (max_length=1 with a 2-unit surrogate pair under
utf16_len): the chunk intentionally exceeds max_length by that one
indivisible codepoint, because emitting it whole preserves content
where the alternatives are data loss or an infinite loop.
Tests: from_json normalizes 0 and negative bounds to 4096 and passes a
real positive bound through unchanged; the sub-codepoint budget emits
whole codepoints with no data loss (all emojis preserved) and a chunk
that necessarily exceeds the 1-unit budget.
BasePlatformAdapter.truncate_message() splits an over-length reply into
chunks. When max_length is 0 or 1 (and the content is longer), the split
loop makes no progress and spins forever, appending empty chunks — an
unbounded hang that pins a CPU and grows the chunk list until OOM:
- headroom = max_length - INDICATOR_RESERVE - ... goes negative, and the
< 1 fallback (max_length // 2) is also 0;
- so _cp_limit is 0, the region is empty, no split point is found, and
split_at falls back to _cp_limit (0);
- chunk_body is remaining[:0] = "", remaining never shrinks, loop repeats.
The same stall is reachable under utf16_len (Telegram) whenever the next
char is a surrogate-pair emoji wider than the whole budget, so _cp_limit
maps to 0 codepoints even for max_length >= 2.
A pathological max_length is not hypothetical: the relay capability
descriptor's max_message_length is taken verbatim from the connector
(gateway/relay/descriptor.py from_json) and assigned straight to the
adapter's MAX_MESSAGE_LENGTH (gateway/relay/adapter.py), and 0 is a
documented "no limit" value there.
Guarantee forward progress: floor headroom at 1, and floor the
final split_at at max(1, _cp_limit) so at least one codepoint is always
consumed per iteration. Normal splitting is unaffected (both floors only
bite when the budget is already degenerate).
Adds regression tests that run truncate_message on a worker thread and
fail if it doesn't return: max_length 0/1/2 terminate and preserve every
character, and the utf16 emoji case terminates too.
Follow-up to the salvaged #65636: if the dedup UPDATE or the retried
CREATE INDEX raises, log and continue — the unique title index is an
optimization and must not block SessionDB initialization.
A fallback chain entry can name its API key via key_env (or the
api_key_env alias) per the fallback-providers docs, but only the gateway
path resolved it — TUI/desktop, cron, and CLI setup fallbacks ignored it,
so a fallback provider whose key lives in a non-standard env var never
resolved on those surfaces.
Centralize the inline-api_key-then-key_env lookup in
hermes_cli/fallback_config.resolve_entry_api_key() and use it at all four
fallback resolution sites (tui_gateway, cron scheduler, gateway runner,
CLI setup mixin); the CLI mixin also gains the base_url passthrough the
other surfaces already had.
Salvaged from PR #43861 (surgical reapply — the original branch predates
the #65264 fallback restructuring).
Rebuilt from PR #61283 onto the /p/<profile>/ routing world (7aa21e336):
_profile_scope(None) now enters the DEFAULT profile's runtime scope when
multiplexing is active instead of returning nullcontext(). api_server is
a port-binding platform living on the default profile, so plain requests
(no /p/ prefix) are the primary path — with fail-closed get_secret they
crashed with UnscopedSecretError on the first credential read (#61276).
All three wrapped call sites (chat-completions executor, /v1/runs agent
construction and _run_sync) inherit the fix through the one seam.
Single-profile gateways keep the no-op. Regression tests ported from
the original PR to the _profile_scope seam.
Fixes#61276
Subset of PR #61985: _make_adapter_auth_check gains a profile_name
parameter and secondary-profile adapters (started in
_start_one_profile_adapters) bind it, so the auth callback's
SessionSource resolves the routed profile's adapter and pairing store
instead of silently falling back to the default profile. This is the
gap left open by the #65629 merge — adapter-internal auth checks (e.g.
Slack thread-context fetch) fire outside the wrapped message handler.
The PR's authz_mixin.py hunks are dropped: main's _auth_env (merged via
PR #65629) already covers the scoped allowlist reads they targeted.
With the mode-conditional check centralized, default (websocket) Feishu
no longer counts as port-binding in the secondary batch report — pin the
fixture to connection_mode=webhook so the test still exercises the
multi-platform report path.
Feishu was unconditionally listed in _PORT_BINDING_PLATFORM_VALUES,
causing the multiplexer to reject ALL Feishu secondary profiles. But
Feishu in websocket mode (the default) uses an outbound WebSocket
connection and does NOT bind an HTTP port — only webhook/callback mode
needs a listener.
Add _platform_binds_port() helper that checks connection-dependent
platforms (currently only Feishu) against their actual config before
raising MultiplexConfigError. Feishu websocket profiles are now allowed;
Feishu webhook profiles still raise as before.
Fixes#52563
Review follow-up: honor source.profile and enter _profile_runtime_scope
only when gateway.multiplex_profiles is on, mirroring the gating in
_run_agent, _reset_notice_session_info, and _resolve_profile_for_key.
When multiplexing is off (the default) a stamped source is ignored and
/profile reports the active profile and default home, byte-identical to
before this PR.
The stamped-source test now enables multiplexing (it previously
exercised the ungated path under the default config), and a new
regression asserts the stamp is ignored when multiplexing is off.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
On a multiplexed gateway the process-level active profile is always the
multiplexer's own (usually "default"), so /profile answered "default" in
every chat regardless of which profile actually served it — making
per-chat persona routing look broken when it was working.
Report source.profile (stamped by the /p/<profile>/ URL prefix, a
per-credential adapter, or a room->profile map) and resolve the
displayed home under that profile's runtime scope, mirroring the scoped
/reset banner (#59003). Unstamped sources fall back to the active
profile and default home, so single-profile gateways are unchanged.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The #62803 branch predates PR #64636's Telegram token-shape validation
on the messaging platform PUT endpoint; align the new guard test's
fixture with the validated format.
The Channels API (PUT /api/messaging/platforms/{id}) accepted and persisted
enabling a port-binding platform on a secondary profile while
gateway.multiplex_profiles is on — a config the gateway only rejects on its
next start, aborting startup with MultiplexConfigError for every multiplexed
profile.
Validate before any .env/config.yaml write and return 409 for the enable
attempt. Disabling and clearing env stay allowed so an already-invalid
profile can be repaired. The port-binding platform set moves to
gateway/config.py (PORT_BINDING_PLATFORM_VALUES) as the single source of
truth shared by gateway startup validation and the dashboard, so the two
policies cannot drift. Platform config mutations now get a names-only audit
log line.
Fixes#62791
_wait_for_callback still read the legacy module-level _oauth_port, so
with two concurrent OAuth flows, flow A's callback wait bound flow B's
port while A's redirect URI pointed at A's port — the callback-side
half of the cross-flow collision that #65622 fixed on the redirect
side. _make_callback_waiter(port) closes over each flow's resolved
port; both provider construction sites (build_oauth_auth and
MCPOAuthManager._build_provider) now wire per-flow waiters. The legacy
_wait_for_callback delegates for backwards compatibility.
Direction credit to @LeonSGP43 (#34280) and the #34260 analysis.
_wait_for_callback catches OSError on bind with a comment claiming the port is
held by a server build_oauth_auth started, and promising to fall back to polling
it. build_oauth_auth never starts a callback server (this is the only listener),
so there is nothing to poll: the branch just raised a misleading "OAuth callback
timed out" when the real cause is a busy port (a concurrent login, a leftover
listener, or a fixed oauth.redirect_port that collided).
Fix the stale comment and raise an accurate, actionable message (names the port,
suggests freeing it or setting a free oauth.redirect_port), chained from the
original OSError. Behavior is otherwise unchanged. Adds a regression test.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A bare False from connect() on EADDRINUSE made the gateway reconnect
watcher treat a port conflict as transient and retry forever at the
backoff cap — 1568+ retries over 5 days in a multi-profile production
setup, filling errors.log and leaking 2 ResponseStore fds per retry.
Set a non-retryable fatal error (api_server_port_in_use) in the bind
OSError branch so the platform drops from the reconnect queue;
recover via /platform resume api_server after changing the port.
Re-implementation of #52132 by @msalles1 against the direct-bind path
from #65621 (the pre-probe block their patch targeted no longer
exists). Their production diagnosis and test scenario preserved.
_clean_reasoning_effort kept its own whitelist that stopped at 'max',
silently dropping 'ultra' from MoA slot configs. Route it through
hermes_constants.parse_reasoning_effort — the same one-source-of-truth
fix the salvaged commit applies to the gateway — so future effort
levels can't drift here either. Docs updated to list ultra.
Follow-up to salvaged PR #64012.
Only fallback resolution needs an explicit target model. Keep the primary resolver call compatible with existing callers and test doubles while retaining atomic provider/model fallback selection.
Switch provider and model together after setup-time auth failure. Serialize global auth-store merges under target-specific locks and preserve auth-to-shared lock ordering for profile OAuth refreshes.
Adds the proxied-callback option to the remote/headless OAuth section,
links the mcp-oauth-remote-gateway skill for fully headless gateways,
and documents the WAF pitfall behind redirect_host.
- redirect_host tests: localhost swap, precedence of full redirect_uri,
empty-value fallback, client-metadata propagation
- The salvaged redirect-uri hint tests predate the #57836
OAuthNonInteractiveError guard; monkeypatch _is_interactive like the
sibling tests in TestRedirectHandlerSshHint
Reclaim.ai's AWS API Gateway WAF 403s any /oauth2/authorize request whose
query string contains a literal 127.0.0.1, so the SDK's hardcoded
redirect_uri made the browser flow impossible. New optional oauth config
key redirect_host (default 127.0.0.1, unchanged behavior) lets a server
entry use localhost instead.
Integrated into _resolve_redirect_uri so it composes with redirect_uri:
an explicit redirect_uri wins; redirect_host only rewrites the loopback
default's hostname.
The PR added a configurable `redirect_uri` (proxy/Funnel callbacks) but
shipped without tests, and the loopback SSH-tunnel hint stayed hardcoded —
actively misleading the exact proxy user the feature targets.
- Extract `_resolve_redirect_uri(cfg, port)` so the client-metadata and
pre-registration paths derive an identical callback (a mismatch makes the
authorization server reject the redirect).
- Make `_redirect_handler` redirect_uri-aware: a configured proxy callback
reaches this machine on its own, so it no longer prints the `ssh -N -L`
loopback guidance. Wired via `functools.partial` — no new global state.
- Document `redirect_uri` in the config block.
- 14 new tests (red/green TDD): helper resolution + empty-string fallback,
metadata + pre-registration for configured/default, AnyUrl normalization,
no-client_id skip, client_secret combo, and both SSH-hint branches.
ruff clean · 81 passed (tests/tools/test_mcp_oauth.py) · ty baseline unchanged
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Widens the deepseek-v4-flash addition to the whole stale-snapshot class:
- deepseek-v4-pro: $1.74/$3.48 → $0.435/$0.87, cache-read $0.003625
(DeepSeek's 2026-07 price cut; every pro session was over-reporting 4x)
- deepseek-chat / deepseek-reasoner: deprecated 2026-07-24, now alias
v4-flash non-thinking/thinking modes — repriced to match flash
(reasoner was $0.55/$2.19 with no cache rate)
- cache_read added to every row; pricing_version unified at
deepseek-pricing-2026-07
- invariant tests: aliases price identically to flash; every deepseek
row carries cache_read < input
DeepSeek's /models endpoint returns no pricing, so direct-provider routes fall back to the _OFFICIAL_DOCS_PRICING snapshot. The table included deepseek-v4-pro but not the newer deepseek-v4-flash, so flash sessions reported $0.00 with cost_source "none". Add the flash entry (values from DeepSeek's official pricing page, mirroring the v4-pro entry; DeepSeek bills no separate cache-write cost) plus two regression tests.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Follow-ups on top of the cherry-picked cluster commits:
- slack: scope-authoritative app-token read — get_secret() with a
narrow UnscopedSecretError fallback to os.getenv. Keeps @kohoj's
correct semantics (scoped profile can never silently inherit the
default profile's Socket Mode app) while fixing the regression where
the default-profile startup loop and background reconnect rebuild,
which call connect() unscoped under multiplex, would raise and
fail-loop. Supersedes the 'or os.getenv' variant from #64461 which
reintroduced the cross-profile fallback leak.
- test: unscoped-multiplex fallback regression test for connect().
- run.py: convert the last legacy self.adapters.get(source.platform)
site (_rename_discord_auto_thread) to _adapter_for_source(source)
so profile-routed Discord sources rename threads on the right
adapter (from #57417's sweep).
- AUTHOR_MAP entry for @aguung.
Partial cherry-pick of a7ffbbff7 from PR #63256: secondary-profile
adapter creation errors no longer abort the whole secondary startup
(try/except around _create_adapter + loud warning on None return), and
Home Assistant's check_ha_requirements() becomes dep-only with the
credential moved to a new validate_ha_config() so secondary profiles
whose HASS_TOKEN lives in the profile secret scope are not silently
dropped by the registry gate.
Telegram diagnostic hunks and profile-label stamping dropped: the
regression they targeted does not exist on current main and they
conflict with the connect() teardown fence.
Per egilewski's security review, WEIXIN_BASE_URL and WEIXIN_CDN_BASE_URL
were still resolved from process-global environment variables, leaving
mixed-scope bypasses in multiplex mode.
Changed files:
- gateway/platforms/weixin.py: Added get_secret import, replaced os.getenv()
with get_secret() for WEIXIN_ACCOUNT_ID, WEIXIN_TOKEN, WEIXIN_BASE_URL,
WEIXIN_CDN_BASE_URL in WeixinAdapter.__init__() and send_weixin_direct()
- tools/send_message_tool.py: Added get_secret import, replaced os.getenv()
with get_secret() for all WEIXIN_* fallbacks in _handle_send()
All runtime Weixin send paths now resolve both credentials and endpoint
configuration from the same profile-scoped source.
Secondary profiles under gateway multiplex keep tokens/allowlists in
profile secret_scope, not process os.environ. Auth and Slack were still
reading os.getenv, so Slack on a secondary profile failed allowlist and
socket mode. Webhook deliver also only looked at default adapters.
- Prefer get_secret for allowlists / allow-all flags (authz_mixin)
- Slack app token + allowlist via secret_scope with getenv fallback
- Wrap secondary profile message handlers in _profile_runtime_scope
before auth runs
- Resolve home-channel env from secret_scope / PlatformConfig
- Webhook deliver falls back to _profile_adapters for target platform
- Template key event_type for webhook prompts
_find_free_port() closed its probe socket before HTTPServer re-bound
the port minutes later, leaving a window where another process could
steal it (#22161 by @amathxbt). _reserve_callback_port() now keeps the
selected socket bound (bounded FIFO pool) until _wait_for_callback
adopts it via bind_and_activate=False. Also sets allow_reuse_address
BEFORE binding — the cherry-picked #44872 set it after the constructor
had already bound, where it is a no-op.
Also updates the three #57836 non-interactive-guard tests to the
closure-factory API from #44872.
Two related OAuth fixes:
1. Replace module-level _redirect_handler with _make_redirect_handler()
closure factory that closes over the resolved port. This prevents
cross-server state pollution when multiple MCP servers run OAuth
concurrently (#44588).
2. Set server.allow_reuse_address = True on the ephemeral callback
HTTPServer so the socket doesn't stay in TIME_WAIT after the flow
completes. This prevents 'Address already in use' errors on the
next OAuth flow for the same port (#44590).
Fixes#44588Fixes#44590
The single-family pre-probe (_port_is_available) raced the real bind and
reported a lingering TIME_WAIT socket as 'in use', failing gateway
restarts for up to ~60s (#10297). Port the webhook adapter's bind
mechanics (#63711/#65482): delete the probe, bind directly with clean
OSError handling and runner teardown, and scope reuse_address=False to
macOS only so Linux restarts rebind past TIME_WAIT instantly.
Credit to @lrawnsley (#10297) for identifying the TIME_WAIT restart
failure.
The overview's total_input/output/cache token counts summed only the
sessions counters (main-loop usage), while the per-model breakdown
already included auxiliary usage rows (task dimension from #65537) and
reconciled residuals. Result: hermes insights top-line totals
undercounted aux spend (compression summarizer, vision, titles) and
disagreed with the per-model table below them — the symptom reported
in #58592 and requested in #9979.
When the per-model breakdown is available, derive the overview token
totals from it (same pattern total_cost already used). Verified no
double-count across incremental CLI deltas, gateway absolute
overwrites, and aux rows.
Guard staged node-pty ASAR path rewrites so already-unpacked paths are
not rewritten twice. Normalize spawn-helper to mode 0755 in both the
prebuild and locally compiled build/Release staging paths.
Add behavioral coverage for both unpacked path forms and both helper
layouts.
Co-authored-by: zhouwei <zwcf5200@163.com>
Co-authored-by: liuhao1024 <sunsky.lau@gmail.com>
Two turns interleaving on one session corrupt the durable transcript:
flushes race (user rows persist out of arrival order), the identity-marker
dedup over shared history dicts can swallow a row, and the second turn
runs on a history base that never saw the first turn's exchange. The
dispatch route that lets the second turn through the busy guard is not
yet identified.
Add note_turn_start (build_turn_context) / note_turn_persisted
(_persist_session funnel): one WARNING naming both turn_ids when a turn
starts before the previous turn's turn-end persist. Ownership transfer
keeps a crashed turn from warning more than once; the unconditional clear
makes the tripwire under-report rather than double-report under a real
overlap. Log-only, no behavior change.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Add a telegram.free_response_topics config list of '<chat_id>:<thread_id>'
entries (plus the TELEGRAM_FREE_RESPONSE_TOPICS env bridge) so a single
forum topic can be free-response — the bot replies without a mention —
without opening the whole chat via free_response_chats. A missing
message_thread_id is normalized to the General topic ('1') via
_effective_message_thread_id.
Re-ported from PR #36049 (by @wesleion): the original patched
gateway/platforms/telegram.py, which has since moved to
plugins/platforms/telegram/adapter.py with a second gating site
(_should_observe_unmentioned_group_message) and plugin-hook config
bridging (_apply_yaml_config). Both gating sites now honor
free_response_topics.
Salvaged-from: #36049
PR #36051's values went stale since May 31: session-store SCHEMA_VERSION
is now 21 (PR said 14), and the dashboard ships 8 built-in themes
(PR said 7). Also document the v16/v18/v20 data migrations added since.
Cross-checked website/docs against the source at main HEAD and corrected
documented commands, env vars, config keys, headers, and default values
that don't match the code. Docs-only; no behavioral changes.
Refs #36048
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
load_transcript is now a live-replay restore site that heals alternation
violations on load (#64934), so the old all-user 120-row fixture was
merged into a single message and the precondition len==120 failed. The
fixture was never a valid conversation shape; alternate user/assistant
so it exercises the same bloat scenario without tripping the repair.
A turn that persists a user row with no assistant row (suppressed reply,
or two concurrent turns interleaving their flushes) leaves a user;user
pair in state.db. The defensive pre-request repair_message_sequence then
re-fires on EVERY request for the rest of the session's life — it mutates
only the per-request list, never the stored transcript.
Add repair_alternation (default False) to get_messages_as_conversation
and pass it from the three live-replay restore sites (gateway
load_transcript, CLI session resume x2). Inspection/export consumers
(trace upload, context guard, api_server history) keep the verbatim
default.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Telegram only auto-derives a voice/audio clip's duration from container
metadata for short recordings; clips longer than ~4:50 are delivered with
duration 0 and render as 0:00 in the player. Probe the length locally
(stdlib wave -> mutagen -> ffprobe) and pass duration explicitly to
sendVoice/sendAudio. Best-effort: when nothing can read the file we omit
duration and fall back to Telegram's prior behavior.
Extracts and hardens the Telegram-only part of the stale, Piper-bundled
PR #7815 (ffprobe-only, predates the send_voice retry/anchor refactor);
relates to #8508.
When an agent is running, the gateway runner's running-agent block routes
"session-level toggles that are safe to run mid-agent" through a membership
set before the catch-all that rejects everything else with
"Agent is running — /<cmd> can't run mid-turn".
A dedicated /footer dispatch branch already sat inside that guard, but the
set listed only {"yolo", "verbose"}, so the footer branch was unreachable:
/footer fell through to the catch-all and was rejected, forcing users to
/stop a running agent just to toggle the runtime-metadata footer. /footer
is a pure display toggle like its sibling /verbose — it only writes
display.runtime_footer.enabled and returns a status string — so it belongs
in the same set.
Add "footer" to the set so the existing dispatch branch becomes reachable.
Regression test (tests/gateway/test_footer_command_mid_run.py): asserts
/footer and /footer <arg> dispatch to _handle_footer_command while an agent
is running, with a /verbose parity guard. Verified failing before the fix
(handler awaited 0 times) and passing after.
Follow-up on the salvaged #64611 commit: the original guard blocked the
install tree unconditionally, which would have broken the legitimate
'developing Hermes from a source clone' CLI flow (launching hermes inside
the repo and getting its AGENTS.md as project context).
Refined policy:
- resolve_context_cwd(): validates configured paths (missing dir -> None +
warning) but honors an EXPLICIT install-tree cwd verbatim — deliberate
user choice.
- build_context_files_prompt(): blocks only the cwd=None -> os.getcwd()
FALLBACK into the install tree, with a new allow_install_tree_fallback
param. system_prompt.py passes it for platform cli/tui (launch dir is
the user's real shell cwd there); desktop/gateway surfaces keep the
guard (their fallback dir is self-spawned, never user-picked).
- Warning log names the resolved dir and the terminal.cwd remedy.
E2E-verified all five scenarios: desktop fallback blocked, in-tree CLI dev
keeps AGENTS.md, explicit install-tree cwd honored, invalid TERMINAL_CWD
falls to None then blocked, normal workspace loads.
_extract_parallel_scope_path used Path.cwd() (process cwd) instead of the
tool's actual execution cwd, and os.path.abspath() instead of os.path.realpath(),
so symlink aliases and relative/absolute path pairs that resolve to the same
physical file were treated as distinct targets and placed in the same parallel
segment. On case-insensitive platforms (Windows) os.path.normcase() was also
absent, allowing Foo.txt and foo.txt to race.
Changes:
- agent/tool_dispatch_helpers.py: introduce _canonical_path(raw_path,
execution_cwd) applying expanduser->abspath->realpath->normcase; thread
execution_cwd through _extract_parallel_scope_path and
_plan_tool_batch_segments
- agent/tool_executor.py: pass get_active_env(effective_task_id).cwd as
execution_cwd to _plan_tool_batch_segments; add pathlib.Path import
- run_agent.py: pass active env cwd to _plan_tool_batch_segments at the
second call site inside _execute_tool_calls
- tests/run_agent/test_tool_batch_segmentation.py: add 5 regression tests
covering relative/absolute same target, symlink alias, execution_cwd vs
process cwd, symlink parent + nonexistent write target, and Windows
case-insensitive alias (skipped on non-Windows)
Fixes a file-corruption / lost-update race introduced by the mixed
tool-batch segmentation feature (perf commit #64460).
Follow-ups to @SAMBAS123's #64986 salvage:
- Replace the hardcoded token-platform set in _platform_has_bot_credential
with PLATFORM_TOKEN_ENV_NAMES, a shared canonical map in gateway/config.py
also used by the empty-token validation warning — one source of truth, so
future token platforms can't silently bypass the gate or drift between
the two sites.
- After secondary-profile startup, warn loudly for any platform skipped on
the primary that no secondary profile ended up serving: an enabled
platform with no credential anywhere is a config error, not a silent
no-op.
- AUTHOR_MAP entry for the salvaged commit's author email.
When gateway.multiplex_profiles is on, the default-profile GatewayRunner
used to call load_gateway_config() unscoped. Platform tokens that lived
only in a profile .env (often a secondary profile) never reached the
primary Telegram adapter, producing "No bot token configured" and an
infinite reconnect watcher loop (#64674).
- Load primary config under the default profile secret scope when multiplex
is enabled (same path secondary adapters already use).
- Skip starting token platforms on the default profile when no credential
is present under multiplex; secondary profiles still connect with their
scoped tokens.
- Drop empty-token configs from the reconnect queue so they cannot spin
forever.
Regression coverage in tests/gateway/test_64674_multiplex_primary_token_scope.py.
Extends @AlexFucuson9's 3-site fix (#58594) across the full adapter:
every logger call and SendResult.error that interpolates a raw PTB
exception now routes through _redact_telegram_error_text(). Covers
polling conflict/retry/network ladders, overflow-split edits, draft
sends, prompt/approval/clarify/picker sends, media send fallbacks,
media cache failures, reactions, and chat-info lookups (48 additional
sites). Telegram Bot API exceptions embed the token in the request URL
(/bot<TOKEN>/<method>), so any raw str(exc) is a leak surface.
Adds regression tests for SendResult.error redaction (update prompt,
clarify) and delete_message debug-log redaction.
Telegram Bot API URLs carry credentials in the path as
/bot<TOKEN>/<method>. Three error-handling paths logged raw exception
text that could include these URLs:
- sendRichMessage fallback (line 1603)
- editMessageText fallback (line 1709)
- polling reconnect warning (line 1902)
Replace raw / with which
uses the existing redact_sensitive_text(force=True) pipeline. This
matches the pattern already used by transient send failures, retry
errors, and legacy edit paths.
Fixes#58376
mark_awaiting_text is the 'Other (type answer)' mode-flip, not a send-time
setup call — invoking it in send_clarify forces the user's next message to
be captured as the clarify response, racing the button-click path and
bypassing the buttons entirely. Telegram calls it only in the 'other'
callback branch; do the same here.
test_run_prompt_submit_requeues_all_unstarted_notifications_with_real_threading
asserted strict FIFO order of the requeued events. The completion_queue is
process-global, and notification pollers leaked by earlier session.init tests
in the same file legitimately steal-and-requeue foreign-session events
(_notification_poller_loop's belongs-elsewhere branch), rotating the queue —
CI slice 6/8 saw [batch_3, batch_2] and failed on ordering alone.
Assert the actual requeue contract instead: batch_1 is consumed while
batch_2 and batch_3 both remain queued — membership via a deadline drain
(an event can be transiently held by a poller mid-cycle), not order.
Verified: single test 10/10 green, full file 337/337 across 3 consecutive
CI-parity runs.
- Refresh _OFFICIAL_DOCS_PRICING fireworks entries against current
docs.fireworks.ai/serverless/pricing: qwen3p6-plus is gone (replaced
by qwen3p7-plus); add glm-5p2/5p1, kimi-k2p7-code, deepseek-v4-flash,
minimax-m3/m2p7, gpt-oss-120b/20b, and the routers/*-fast tiers with
their distinct higher rates.
- Picker pricing via get_pricing_for_provider('fireworks'): pure dict
transform over the shared models.dev in-memory/disk cache (1h TTL) +
_pricing_cache memoization — no new network call on the picker path.
- Wire pricing display into the generic api-key-provider setup flow so
Fireworks model pickers show $/M columns like OpenRouter/Nous do.
- Invariant tests: plugin fallback_models all priced, fast tiers price
higher than standard, every row carries cache_read < input.
Fireworks-hosted sessions previously showed estimated_cost_usd = 0
because (a) _OFFICIAL_DOCS_PRICING had no Fireworks entries and (b)
resolve_billing_route() had no branch for provider="fireworks",
falling through to billing_mode="unknown".
Adds entries for the three Fireworks models hermes operators are
most likely to route through (Kimi K2.6, DeepSeek V4 Pro, Qwen3.6-Plus)
and a routing branch that triggers on either explicit
provider="fireworks" or api.fireworks.ai base_url match. Mirrors the
recently-merged MiniMax addition pattern; pricing snapshot sourced
from https://docs.fireworks.ai/serverless/pricing and the per-model
pages on fireworks.ai.
Tests cover: (a) full Fireworks model id resolves to the snapshot
entry, (b) base_url alone is sufficient to route, (c) end-to-end
estimate returns "estimated" status with the expected dollar amount.
A follow-up upstream issue is open proposing a dynamic pricing
source (e.g. litellm's pricing JSON) as a permanent fix to the
PR-per-model treadmill that this snapshot keeps adding to.
_clamp_duration returned durations[0] for all families when duration=None,
causing pixverse-v6, seedance-2.0, and kling-v3-4k to always send their
minimum value (1s, 4s, 3s respectively) instead of omitting the field and
letting the FAL endpoint apply its own default.
Range families are now detected via the existing _is_duration_range
heuristic and return None (field omitted) when no duration is requested.
Enum families like veo3.1 keep sending their first entry as the default.
Follow-up for salvaged PRs #56055 + #56803, adapted to the per-session
cwd record store (PR #65213): the record (live cd state) is rung 1,
the registered session.cwd.set override rung 2, TERMINAL_CWD rung 3 —
the same ladder file tools and the terminal resolve against. Covers
record-over-override, record-only, and stale-record fall-through.
execute_code's _resolve_child_cwd() only checked the process-global
TERMINAL_CWD env var and os.getcwd(), ignoring the per-session cwd
override registered via session.cwd.set → register_task_env_overrides.
This caused execute_code to write to the process launch directory
while sibling tools (write_file, read_file, patch, terminal) correctly
resolved the session workspace — two file-writing paths in one turn
silently disagreed on the working directory.
Fix: pass task_id to _resolve_child_cwd() and check
_registered_task_cwd_override(task_id) before falling back to
TERMINAL_CWD and os.getcwd(), matching the lookup order used by
file_tools._resolve_base_dir and terminal_tool._resolve_command_cwd.
MEDIA: tags whose path had an unknown extension (.py, .log, .toml,
.weirdext, ...) fell between both extraction passes: the anchored
extension allowlist (MEDIA_TAG_CLEANUP_RE) did not match them, and the
extension-less pass explicitly skipped any path that HAD a suffix. The
file was never delivered even though the intended design (universal
ingress/egress) says any non-credential file should ship.
Widen _path_lacks_deliverable_extension() so the validated delivery
pass (MEDIA_EXTENSIONLESS_TAG_RE + validate_media_delivery_path)
covers every path the extension allowlist does not — unknown
extensions and extension-less files alike. Security posture is
unchanged: unknown-extension paths only deliver after full validation
(exists, symlinks resolved, credential/system denylist, strict-mode
allowlist+recency), and unvalidated tags stay visible in the text
instead of being silently dropped. Known extensions keep their
unconditional pre-existing behavior.
Because extract_media, _strip_media_tag_directives (non-streaming
dispatch), and strip_media_directives_for_display (streaming) all
share the same two regexes + predicate, all delivery paths pick up the
widened behavior with no per-site changes. Dispatch partition in
gateway/run.py already routes non-image/video extensions through
send_document.
Closes the gap reported in PR #36060; supersedes the allowlist-append
approach there (an extension allowlist can never enumerate every file
type a user asks the agent to produce).
Co-authored-by: Randimt <randimt@users.noreply.github.com>
* feat(analytics): record auxiliary model usage per task in session accounting
Auxiliary LLM calls (vision, compression, title_generation, web_extract,
session_search, ...) discarded their token usage, leaving dashboard
analytics blind to aux model spend (issue #23270).
- hermes_state.py: session_model_usage gains a task PK dimension
(''=main loop) via v22 table-rebuild migration (SQLite can't alter a
PK); record_auxiliary_usage() writes per-(model,provider,task) deltas
WITHOUT touching sessions counters (gateway overwrites those with
absolute main-loop totals — folding aux in would double-count or be
clobbered). Aux rows never inherit the session's main-loop route.
- agent/aux_accounting.py: ContextVar ambient accounting context
(mirrors the portal_tags conversation context); record_aux_usage()
normalizes usage via usage_pricing.normalize_usage, estimates cost,
and is strictly best-effort. moa_reference/moa_aggregator excluded —
conversation_loop already folds MoA usage+cost into the main delta.
- agent/auxiliary_client.py: _validate_llm_response is the recording
chokepoint — every successful non-streaming aux response passes
through it exactly once, sync and async, including fallback paths
(model read from the response itself stays accurate across
fallbacks).
- run_agent.py: run_conversation publishes/resets the accounting
context; agent/title_generator.py republishes on its bare thread.
- hermes_cli/web_server.py: /api/analytics/usage folds aux rows into
by_model (aux-only models finally appear) and adds a by_task
summary; /api/analytics/models surfaces aux rows on the Models page.
Design per review of PR #62850 by @eeksock (thread-local + separate
auxiliary_usage table): rebuilt on ContextVar (async-safe — thread-local
cross-attributes concurrent coroutines on one event loop) and the
existing session_model_usage table instead of a parallel accounting
path, extended beyond vision to every aux task, and wired the analytics
endpoints so the dashboard actually shows it. Credit to @eeksock for
the approach and @tboatman for the detailed root-cause analysis.
* test(moa): match _validate_llm_response mock to new accounting-hint signature
* test(aux): accept accounting-hint kwargs in remaining _validate_llm_response mocks
Docs and MultiplexConfigError already promise secondary profiles are
served through the shared listener's /p/<profile>/ prefix, but only the
webhook adapter registered those routes — api_server returned 404.
Mirror every HTTP route, validate the prefix, and scope agent runs /
session DB / model listing to the target profile.
Add an optional skill for connecting OAuth-gated remote MCP servers
(Better Stack, Linear, Cloudflare, Datadog, Stripe, etc.) when Hermes
runs as a remote gateway, where the built-in browser OAuth flow cannot
capture the 127.0.0.1 callback.
Covers the manual RFC 7591 DCR + RFC 7636 PKCE + authorization_code
flow, writing tokens in Hermes' HermesTokenStorage schema, the
dashboard-first escalation path, and a diagnostic script + pitfalls
for refresh/session-revocation recovery.
Follow-up for salvaged PR #63255: with LocalEnvironment._update_cwd
delegating to the stdout marker parser, the cwd temp file has zero
readers left. Drop the 'pwd -P > file' writes from the bootstrap and
_wrap_command so every command stops paying a pointless file write
(and stops littering temp dirs with hermes-cwd-*.txt).
On Linux, SO_REUSEADDR only allows rebinding past TIME_WAIT (a second
live listener would need SO_REUSEPORT, which we never set), so
disabling it bought no protection there while making a quick gateway
restart fail to rebind for up to ~60s. Keep the BSD silent-split guard
on darwin, default semantics elsewhere.
E2E verified: dual-stack v4+v6 bind on one port, immediate rebind
after disconnect, and live-listener conflict still rejected.
Follow-up for salvaged PR #26790: on a POSIX host with _IS_WINDOWS
patched (test simulation), os.path.isabs rejects C:\Users\x and the
new relative-cwd recovery would mangle a perfectly absolute native
Windows path. Check ntpath.isabs first on the Windows branch.
tools_disable_enable_command filters platform-restricted toolsets out of
toolset_targets and prints an error for each, but the success summary at
the end is built from the raw targets list and only excludes unknown
toolsets and failed MCP servers. Running e.g.
hermes tools enable discord --platform telegram
prints the 'not available on platform' error followed by 'Enabled:
discord' for a toolset that was never written to the config.
Exclude restricted_targets from the success summary, matching how
unknown toolsets and failed MCP servers are already handled.
Two regression tests: a restricted toolset alone must not print
'Enabled', and a mixed allowed+restricted invocation must report only
the allowed toolset (both fail before the fix).
Extends the conversation=<id> Portal tag (salvaged from PR #65183 by
@J-SUPHA) from main-loop-only to every LLM call in a conversation:
- agent/portal_tags.py: ContextVar-based conversation context.
nous_portal_tags() falls back to the ambient id when no explicit
session_id is passed, so every aux tag site (auxiliary_client,
chat_completion_helpers summary path, web_tools) inherits the tag
with zero per-call-site plumbing. Ambient id wins over explicit
per-segment ids since it carries the lineage root.
- hermes_state.py: SessionDB.get_conversation_root() — public wrapper
over the lineage walk; returns the ROOT session id, so one
user-facing conversation keeps a single conversation= value across
context-compression rotation, and delegate subagent trees tag as
their parent conversation.
- run_agent.py: run_conversation() publishes the root id for the turn
and resets it in finally. _conversation_root_id() resolves via
_parent_session_id for subagents.
- agent/moa_loop.py: MoA reference fan-out workers now run under
propagate_context_to_thread so advisor slots attribute to the acting
conversation (also fixes approval-callback propagation on that path).
- agent/title_generator.py: bare title thread republishes the context
from its session id (spawned after turn reset).
Tests: ContextVar semantics, cross-context isolation, thread-hop
propagation, lineage-root resolution incl. cycle guard.
The end-to-end _build_api_kwargs parity test asserted the Nous Portal
tags exactly equal the base two-tag list. With the per-session
conversation tag, a real agent (which has a session_id) now emits a
third `conversation=<session_id>` tag. Assert against
nous_portal_tags(session_id=agent.session_id) so the check stays exact.
Widen @lEWFkRAD's sidecar-headless fix (PR #54565) to the sibling spawn
sites: the npm ci / npm install self-heal runs in _reinstall_sidecar_deps
also popped a brief console window per run on Windows. Same
windows_hide_flags() helper (CREATE_NO_WINDOW only, so capture_output
stays usable).
plugins/platforms/photon/adapter.py launches the Node sidecar (and the
spectrum-ts mixed-attachment patch run) via subprocess without creationflags.
On Windows this opens a visible console window on every sidecar (re)start --
and because a failed sidecar is retried on a timer, it flashes repeatedly.
Wire windows_hide_flags() (hermes_cli/_subprocess_compat) into both spawns,
the same helper the discord and whatsapp adapters already use for their
sidecar spawns -- photon was the one platform adapter this pattern missed.
CREATE_NO_WINDOW only (no DETACHED_PROCESS) so the persistent sidecar's
stdin/stdout pipes stay usable for the supervisor.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The per-session record store is now the ONLY cwd mechanism. Deleted:
- env.cwd_owner stamping + prev_owner threading (terminal_tool): the
shared env no longer carries ownership metadata at all
- _resolve_command_cwd's env/prev_owner params: resolution is
workdir > session record > config/override default
- file_tools._live_cwd_if_owned + _get_live_tracking_cwd: path
resolution never consults the shared env's live cwd
- file_tools._last_known_cwd + _remember_last_known_cwd +
_last_known_cwd_for: the #26211 preserved-anchor registry is
subsumed by the session record, which never lived on the env and
therefore cannot be lost to env cleanup. The _get_file_ops
stale-cache rescue now writes the record instead.
- env recreation (both _get_file_ops and terminal_tool) seeds the
fresh env from override > session record > config
Why no transition fallback: the legacy state was process-local and
in-memory exactly like the record store — after a restart both start
empty, and within a running process every legacy write site has been
dual-writing the record since step 1. There is no populated-legacy/
empty-record state to fall back for.
Tests updated to drive the record store instead of the deleted
mechanism; the cross-session isolation suite now asserts the same
behavior contracts (no leak, cd isolation, #26211 persistence)
against the new architecture, plus a new "session C inherits nothing"
case that the old ownership guard could not express.
Third step of the cwd rearchitecture: _resolve_command_cwd now prefers
the session's own cwd record over the shared env's live cwd.
New resolution order: workdir > session record > legacy env.cwd
(ownership-gated, transition-only) > config/override default.
The record is written after every completed command for the session, so
it IS the session's cd state — another session's cd lands in another
record and cannot affect this session's commands. The legacy env.cwd
branch only fires for a session with no record yet (no command has
completed since this code loaded); it keeps the prev_owner ownership
guard for that transition window and is deleted in step 4 along with
env.cwd_owner stamping and file_tools' _last_known_cwd machinery.
Adds command-path regression tests including the terminal sibling of
the leak-A scenario (unowned shared env cwd vs session record) and an
E2E cd round-trip through terminal_tool.
Flips the read side of the cwd rearchitecture onto the _session_cwd
store introduced in the previous commit.
_authoritative_workspace_root now resolves:
1. the session's own cwd record (get_session_cwd) — per-session by
construction, so one session's cd can never leak into another
session's file resolution, with no ownership heuristics at all
2. registered override (fallback for cleared/never-written records)
3. legacy shared-env live cwd + preserved anchor (transition-only,
for commands that ran before this code loaded)
4. sentinel-free absolute TERMINAL_CWD
delegate_task children get their record seeded from the parent's at
spawn: they keep starting in the parent's directory (current behavior)
but their subsequent cds stay isolated in their own record instead of
bleeding back through the shared env.
The wrong-worktree leak class is now solved structurally on this path —
there is no shared cwd for sessions to inherit. The legacy env-side
tracking (cwd_owner, _live_cwd_if_owned, _last_known_cwd) remains only
as a transition fallback and is deleted in the next step.
First step of the cwd rearchitecture (see PR #65185 for the targeted
leak fixes this will eventually supersede, and
.hermes/plans/cwd-rearch-audit.md for the full audit + sequencing).
The root cause of the wrong-worktree bug class is that cwd lives on the
SHARED terminal env — a global mutable timeshared between sessions.
env.cwd_owner stamping, _last_known_cwd, and file_tools' ownership
ladder are all patches over that misplacement.
This adds the replacement store: _session_cwd, keyed by the raw
session/task key, with record/get/clear accessors. Step 1 is dual-write
only — every site that learns a session's live cwd also records it:
- terminal_tool foreground path: after env.execute() the env's own
post-command tracking has updated env.cwd; mirror it under the
session key that drove the command
- register_task_env_overrides: a registered workspace cwd (ACP/TUI/
desktop) seeds the session record
- clear_task_env_overrides: drops the record on teardown
Readers are untouched — behavior is identical. Later steps flip
file_tools resolution and _resolve_command_cwd to read this store,
then delete env-side tracking, cwd_owner, and _last_known_cwd.
Also hardens terminal_tool's env acquisition with an explicit
env-is-None guard (previously implicitly unbound on an unreachable
branch, flagged by pyright once the dual-write read env post-loop).
Disable address reuse so an existing family-specific listener cannot silently split traffic with the webhook server. Normalize wildcard bind hosts for local CLI URLs and align setup documentation with the dual-stack default.
The webhook adapter defaulted to host='0.0.0.0' — IPv4 only. On Fly.io
hosted agents the edge router (hermes-agent-router) reverse-proxies public
webhook traffic to <app>.internal:8644 over 6PN, Fly's private network,
which is IPv6-only (.internal resolves to an fdaa:… address). An IPv4-only
listener is unreachable there, so public webhook POSTs to
https://<agent>.agents.nousresearch.com/webhooks/<route> never landed on
the adapter — the router's dial was refused.
Fix: DEFAULT_HOST = None, which makes aiohttp/asyncio create_server bind
BOTH address families. '::' is NOT a valid substitute: on hosts where the
kernel sets bindv6only=1 (verified on Fly machines) it yields an IPv6-only
socket, breaking the IPv4 loopback /health check and the AF_INET
port-conflict probe in connect(). None binds per-family regardless of the
sysctl. An explicit empty-string/null host in config now also normalises to
None (dual-stack) rather than an invalid host=''. Users can still pin a
specific host via platforms.webhook.extra.host.
Validated live on a Fly staging agent: with this default and no config
override, the adapter binds both v4 and v6 (127.0.0.1:8644 and [::1]:8644
both answer), and a public signed webhook POST through the router returns
202 (valid sig) / 401 (bad sig) instead of the router's 502.
Tests: new TestDualStackBind asserts the None default, config resolution
(missing/empty→None, pinned preserved), and a real dual-stack bind opens
both AF_INET and AF_INET6 listeners. Red-proof: these fail on the old
'0.0.0.0' default.
Moves the fireworks entry in CANONICAL_PROVIDERS from its old slot
(after GMI Cloud) to directly below Nous Portal, ahead of OpenRouter.
Order propagates automatically to hermes model, the setup wizard,
Telegram /model, and the desktop provider catalog.
Apply positive-proof routing to every addressed notification in the registry and TUI poller while preserving ownerless legacy behavior and TUI delivery for poll-observed completions.
Remove the unused exact-key drain helper and cover ordinary success and failure, origin, compression-lineage, orphan, and poll-observed paths.
Complements NousResearch/hermes-agent#54785.
Two complementary fixes for cross-session background-process notification
leakage in the TUI/Desktop multi-session path (#42674, #35652).
1. Poller orphan guard: after _notification_event_belongs_elsewhere
returns False, check whether the event has a non-empty session_key
that differs from the current session. If so the owner session is
gone — drop the event instead of hijacking it into an unrelated
session transcript.
2. Post-turn drain filter: the existing drain_notifications() pops every
event from the global queue regardless of ownership. Added
_drain_owned_notifications() which applies the same ownership routing
used by the poller (consume own, requeue foreign-live, drop orphan),
and wired it into the post-turn safety drain.
Complementary to PR #42731 which addresses a separate code path in the
same bug class. Together they close#42674.
git diff on a lockfile is unreadable: npm reorders entries, rewrites
integrity hashes, and moves packages between nesting levels, so a
one-line package.json bump produces a thousand-line textual diff.
scripts/ci/lockfile_diff.py instead parses the `packages` map out of
both versions of every tracked package-lock.json (via `git show`),
reduces each to {install path: version}, and set-diffs the maps —
reorder/hash churn vanishes, leaving only actual version movement
(added / removed / updated, with nested dedup copies tracked
separately).
The lockfile-diff workflow posts the result as a Markdown table in a
PR comment gated behind a hidden marker: subsequent pushes PATCH the
existing comment instead of stacking new ones, and a push that reverts
all lockfile changes updates the comment to say so. Advisory only —
never fails on findings; fork PRs (read-only token) degrade to a
warning.
Wired through the ci.yml orchestrator with a new npm_lock lane in
classify_changes.py (fails open on .github/ changes per the existing
contract).
* fix(nix): fold makeWrapper line continuations into optionalStrings
When rev == null (any dirty-tree build), the empty optionalString
expansion left the previous line's trailing backslash dangling onto a
blank line, ending the makeWrapper command early and running
`--suffix PYTHONPATH ...` as its own shell command (`--suffix: command
not found`, exit 127). Clean trees passed CI; dirty trees with
extraPythonPackages failed — exactly the path the NixOS module
exercises.
The continuation now lives inside each optionalString (" \\\n --set
..."), so the makeWrapper chain stays intact whether or not the
optional flags expand.
Verified by building with rev = null + extraPythonPackages =
[ pyfiglet ]: wrapper builds, PYTHONPATH suffix lands inside the
makeWrapper call, wrapped `hermes --version` runs, and the collision
check still executes (certifi correctly rejected).
* perf(nix): filter derivation sources to shrink rebuild scope
Every derivation previously saw the whole repo, so any file change
rebuilt everything. Each derivation now gets a filtered src with only
the files it consumes:
- lib.nix: derive npm workspace topology from the root package.json
`workspaces` globs (single source of truth — a new workspace member
is picked up with zero nix edits). pythonSrc (cleanSourceWith)
excludes the JS workspace trees, docs/website, docker/.github,
tests, nix/, flake.nix/flake.lock, root docs, and skills/ +
optional-skills/. importNpmLock reads from a fileset-filtered
npmRoot (root manifests + member package.jsons only).
- mkNpmPassthru takes `dirs` — the workspace dirs the package
contains — and builds a per-package fileset src from them. web and
desktop include apps/shared (file: dep). One shared
`nix run .#update-npm-lockfile` replaces the per-package
update_*_lockfile bins that only existed inside build sandboxes.
- python.nix: release venv loads the uv2nix workspace from pythonSrc.
The editable venv keeps an unfiltered ./.. root —
mkEditablePyprojectOverlay calls lib.path.splitRoot, which rejects
a cleanSourceWith set, and the editable install reads the live
checkout anyway.
- hermes-agent.nix: skills ship exclusively via HERMES_BUNDLED_SKILLS
/ HERMES_OPTIONAL_SKILLS (same mechanism as Homebrew packaging;
setup.py's _data_file_tree returns [] for missing dirs), so
SKILL.md edits no longer rebuild the venv. optional-mcps stays in
the wheel — pyproject.toml lists its manifests as explicit
data-files. Bundled assets are symlinked instead of copied, making
the wrapper drv near-instant when only an input changed.
__pycache__ filtered from bundled skills.
- checks.nix: find -L through the new symlinks; assert
optional-skills presence + HERMES_OPTIONAL_SKILLS in the wrapper.
- run_tests.sh: fall back to $HERMES_PYTHON when no local venv
exists, guarded by an `import pytest` probe (HERMES_PYTHON from a
wrapped hermes binary points at the release venv, which has no
pytest — without the guard every test file dies with "No module
named pytest" while the runner exits 0).
Verified: nix flake check exit 0; built .#default .#tui .#web
.#desktop; SKILL.md and flake.nix edits leave the venv drvPath
unchanged; .py edits leave the tui drvPath unchanged; .tsx edits
leave the venv drvPath unchanged (and do change the tui drv);
scripts/run_tests.sh runs 299 tests green through both the venv and
HERMES_PYTHON paths, and rejects a pytest-less HERMES_PYTHON.
* refactor(nix): overlay aliases the flake's own package instead of re-instantiating
The overlay previously re-called callPackage against the consumer's
nixpkgs (final), so pkgs.hermes-agent could be a different derivation
than nix build .#default and the NixOS module's default — an untested
build matrix against arbitrary consumer nixpkgs versions, for a
package whose Python side is uv2nix-locked anyway.
Now the overlay is a pure alias for the flake's own locked package:
one callPackage site (packages.nix), everything else references it.
.override { ... } still works — callPackage's makeOverridable travels
with the derivation.
Verified: direct drvPath == overlaid drvPath; .override produces a
distinct drv.
* fix(nix): dedupe extraPlugins assertions, replace MESSAGING_CWD with terminal.cwd
- Delete the duplicated extraPlugins duplicate-name assertions block
(same assertion declared twice back to back).
- Stop setting the deprecated MESSAGING_CWD env var, which made the
module trip hermes' own startup deprecation warning. The working
directory is now injected as terminal.cwd into the generated
config.yaml; cfg.settings wins via recursiveUpdate, and container
mode maps to the in-container mount path.
hermes-agent is public/OSS; the forensic-logging comment in
_quarantine_nous_oauth_state named 'Fly' (the specific managed-hosting compute
provider) twice. Reword generically ('a hosted agent', 'a managed log drain may
be WARNING-only') — the behaviour is unchanged, only the comment. Follows the
same scrub applied to the boot re-seed helper (#59983) before merge; this one
slipped through in #59976.
When the bot PR auto-merges, main moves — which the poll loop detects
as 'main moved' and tries to close the PR. But the PR is already
merged, so gh pr close fails with an error.
Fix: when main moves, re-check PR state before closing. If it's
already MERGED, exit cleanly. Also make gh pr close non-fatal (|| true)
as a belt-and-suspenders guard against the same race on the other
close paths.
The js-autofix workflow used 'gh pr create --json number --jq .number'
to capture the PR number, but 'gh pr create' doesn't support --json.
Extract the PR number from the URL that 'gh pr create' prints instead.
tests/test_desktop_mac_entitlements.py asserts about
apps/desktop/electron/*.plist — the same CI blind spot as the other
ported tests: the change classifier routes apps/ changes to the
frontend lane, so a PR touching only the plists would skip the Python
suite and the regression would go green on the PR and red on main.
Ported to tests-js/desktop-mac-entitlements.test.ts so it runs in the
correct lane. All three tests carry over: the inherit plist grants
audio-input (regression #37718), every device.* entitlement on the
main app is also inherited, and both files remain well-formed.
Parsing uses the `plist` package (pinned to ^3.1.0, the version
already present in the workspace lockfile, so no new transitive
packages) plus `@types/plist` — Node's stand-in for plistlib.
Verified: tests-js `npm run check` passes (typecheck + 9/9 tests), and
a mutation run (removing audio-input from the inherit plist) turns
both regression tests red.
The CI change classifier routes package.json / package-lock.json /
.ts/.tsx changes to the frontend lane, not the Python lane. Four
Python tests asserted about these JS-side artifacts, so a PR touching
only those files would skip the Python suite — the regression goes
green on the PR and red on main (where the classifier fails open).
Ported all four to vitest so they run in the correct CI lane:
tests/test_package_json_lazy_deps.py
→ apps/desktop/electron/package-json-lazy-deps.test.ts
(camofox is lazy, agent-browser is eager, lockfile clean)
tests/test_desktop_electron_pin.py
→ apps/desktop/electron/desktop-electron-pin.test.ts
(electron dep is exact, matches build.electronVersion, lockfile agrees)
tests/test_assistant_ui_tap_compat.py
→ apps/desktop/electron/assistant-ui-tap-compat.test.ts
(@assistant-ui cluster shares one tap version + semver helper)
tests/test_dashboard_sidecar_close_on_disconnect.py
→ web/src/lib/chat-sidebar-session-params.test.ts
(sidecar session.create opts into close_on_disconnect + profile)
The ChatSidebar test was regex-matching .tsx source text (the
source-reading anti-pattern). Extracted sidecarSessionCreateParams()
from the component's effect into an exported pure function so the
test calls real code instead of pattern-matching a string.
Verified: 8 electron tests + 76 web tests pass; both typecheck clean.
Adds a "Tests for JavaScript / npm / package.json invariants belong
in the JS suite" subsection under Testing, documenting that the CI
classifier routes package.json / lockfile / .ts/.tsx changes to the
frontend lane — so Python tests asserting about those files won't run
on a JS-only PR. Includes a table mapping artifact types to the
correct vitest workspace and run commands.
* fix(js): never format package-lock.json
prettier and eslint should never touch package-lock.json. main has a
repo rule requiring team approval when lockfiles change, so an autofix
PR touching it would hang waiting for review.
- Add .prettierignore at repo root
- Add '**/package-lock.json' to eslint shared config ignores
* fix(ci): js-autofix pushes via PR instead of direct push to main
Main now has repository rules requiring pull requests + required status
checks ("All required checks pass"), so the workflow's direct push to
main is rejected with GH013 every time eslint --fix produces changes.
Switch apply-patch to push to a dedicated bot/js-autofix branch, create
or update a PR, and enable auto-merge (squash). The PR auto-merges 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.
The two-job security split is preserved:
- generate-patch stays unprivileged (contents: read only) — it runs npm
on an ephemeral runner with zero push permissions.
- apply-patch (contents: write + pull-requests: write) still never runs
npm, never installs anything, never executes repo code — it applies
the trusted patch artifact and delivers it via PR.
The hoisted shared eslint config catches pre-existing no-useless-escape
and no-control-regex errors in apps/desktop that were only fixed in web/
in the previous commit.
- no-useless-escape: remove unnecessary \) escapes inside character classes
- no-control-regex: add eslint-disable-next-line comments for intentional
\x1b terminal escape byte matching (same pattern as web/src/lib/pty-mobile-input.ts)
apps/shared had lint:fix but not the 'fix' alias that other workspaces
have. The js-tests check job runs 'npm run fix' as a second step, so
this workspace was failing with 'Missing script: fix'.
- Fix no-useless-escape in i18n/ko.ts and i18n/uk.ts (remove backslash
escapes inside single-quoted strings)
- Add eslint-disable-next-line for no-control-regex in pty-mobile-input.ts
(terminal data legitimately contains control characters)
- Configure react-refresh/only-export-components with allowConstantExport
so context providers that export hooks don't trigger the rule
- Downgrade react-hooks v7 rules (set-state-in-effect, refs,
preserve-manual-memoization, static-components) from error to warn —
these are real concerns but the existing code uses common patterns
(data loading on mount, ref-as-instance-var) that need careful refactoring
Add a 'lint' job to the JS tests workflow that runs 'eslint --fix'
across all discovered npm workspaces (same matrix as the check job).
Fixable issues auto-correct and don't block; eslint exits non-zero
only when un-fixable errors remain.
Also fix duplicate 'needs: workspaces' in the check job.
* feat(desktop): add background-task indicator to sidebar session rows
A session with a live terminal(background=true) process but no active LLM
turn now shows a pulsing gray dot in the sidebar — distinct from the accent
pulse of an active turn and the steady amber/green of needs-input/unread.
New $backgroundRunningSessionIds computed atom joins
$backgroundStatusBySession (runtime-keyed) with $sessionStates
(runtime→stored) to produce stored session ids the sidebar row can match
against — same pattern as $attentionSessionIds and $unreadFinishedSessionIds.
Also refactors SidebarRowDot from a 3-branch if-else chain into a
table-driven priority array of DotState entries. Each state declares its
active flag, className, ariaLabel, and title in one place — adding a new
indicator state is now one array entry instead of another conditional.
* fix(desktop): keep pulse-dot before:bg-* static so Tailwind emits it
The PING() helper interpolated the color into `before:bg-${color}`, but
Tailwind v4 only generates utilities it finds as complete static strings — a
template-composed `before:bg-(--ui-accent)` / `before:bg-muted-foreground/50`
is never emitted, so both pulse rings (working + new background) lost their
halo color. Make PING a static scaffold and write the before:bg color inline
per variant.
---------
Co-authored-by: Brooklyn Nicholson <brooklyn.bb.nicholson@gmail.com>
Thinking-only retries flush the scrubber then stream again without
reset(). flush() was leaving _last_emitted_ended_newline False, so the
next stream's opening <think> looked mid-line and leaked into the UI.
SessionInfo already carries git_branch from the backend, but neither the
ctrl-k command palette nor the sidebar search function included it in
their matching fields. Typing a branch name matched nothing.
- session-search.ts: add session.git_branch to sessionMatchesSearch
- command-palette/index.tsx: thread git_branch through SessionEntry and
into the keywords array for both sessions and archived sessions groups
- session-search.test.ts: cover full, partial, and "main" branch matches
* fix(desktop): un-minimize zone when revealing a collapsed tool panel
`revealTreePane()` fronted the pane's tab but never un-minimized its zone
when the pane lived in a shared zone (terminal + logs, or a tool panel
stacked with the workspace). The shared-zone branch of `setPaneCollapsed`
routes opens through `revealTreePane` instead of `toggleTreeGroupMinimized`,
so the zone stayed collapsed and the terminal appeared to "close but not
open" on Ctrl+` — and tab clicks needed a double-click because the first
click's `restoreTreePane` → opener → listener → `revealTreePane` chain
didn't un-minimize either.
The fix: `revealTreePane` now restores a minimized zone before fronting
the pane, chaining the un-minimize + activate into a single `commit` so
the second op sees the updated tree.
* fix(desktop): side collapse in column-root layouts + restore after zone-menu minimize
Two follow-up fixes for the panel layout:
1. Ctrl+B / Ctrl+J sidebar toggles didn't work in the Terminal deck (and
Quad) layout. The side-collapse system (treeSideOfPane, paneRootSide,
layoutHasRootSide, and the renderer's semanticSides) only operated on the
ROOT split when it was a row — but Terminal deck's root is a column, so the
side columns (sessions/files) nested inside a child row were invisible to
the collapse system. Extracted a rootRow() helper that finds the row
containing main (the root itself for row layouts, or the row child of a
column root), and a rootRow prop propagated through TreeNode → TreeSplit so
semanticSides fires on the right split regardless of root orientation.
2. Clicking the "logs" tab in a minimized terminal/logs zone needed a
double-click when the zone was minimized via the zone menu (right-click →
minimize), not the toggle. restoreTreePane called the opener
($logsOpen.set(true)) and returned early — but when the store was already
true, nanostores don't fire the listener, so setPaneCollapsed/revealTreePane
never ran and the zone stayed minimized. Now restoreTreePane also
un-minimizes directly + calls revealTreePane when the zone is still
minimized after the opener runs.
Follow-up on kshitijk4poor's cherry-picked references:
- pitfalls.md rewritten from the raw-socket blender_exec() frame to the
MCP-tool frame (dropped 'MCP server is optional, talk to the socket
directly' — now the anti-pattern; dropped TCP-helper internals items;
kept all bpy/addon knowledge: empty code results in 5.x, temp-file
readback, ops-vs-data context, engine names by version, GPU setup)
- recipes.md: blender_exec -> execute_blender_code in the agent-side
verification snippet
- SKILL.md: reference-file table added to Quick Reference, version
2.1.0, kshitijk4poor added to authors
- docs page regenerated (scoped to this skill)
Fixes#62452. Two amplifiers turned one slow auxiliary route into a
per-turn multi-minute stall:
1. Fallback candidates inherited the exact effective_timeout the primary
was called with. When the primary's deadline was short (tuned or
already burned), an independently healthy fallback died on the same
clock — the reporter's 163k-token compression needed ~90s on the
fallback and got the primary's 30s, every turn. fallback_chain
entries may now declare their own 'timeout' (seconds); both fallback
candidate call sites (sync + async) resolve it via
_fallback_entry_timeout, label-scoped so only configured-chain
candidates are affected. No entry timeout → task-level timeout,
preserving existing behavior.
2. A session whose transcript structurally cannot be summarized within
the deadline re-attempted every 60s, re-burning the full timeout on
every subsequent turn. Consecutive timeout-class failures now
escalate the cooldown 60s → 300s → 900s (capped); any successful
summary or session reset clears the streak. Timeout classification
takes precedence over the streaming-closed 30s rung ('timed out'
also matches _is_connection_error) and now recognizes the SDK's
'Request timed out.' phrasing.
Fail-safe behavior is unchanged: all messages are preserved when every
candidate fails; the cooldown only spaces out retries.
The served-profiles block in _start_secondary_profile_adapters references
PairingStore, but the class's only import in gateway/run.py is method-local
inside __init__ — so the reference raised NameError at runtime, silently
swallowed by the enclosing try/except ('could not record served_profiles').
Result: multiplexing gateways never created per-profile pairing stores, and
authz pairing checks for secondary profiles fell through to the global
whitelist. Also masked the served_profiles runtime-status write.
One-line fix (local import alongside write_runtime_status) + regression
tests that drive the real method and assert the stores materialize,
verified red without the import and green with it.
Surfaced during the profile-routing sweep by @CocaKova's PR #61689, which
included the same fix as part of a larger feature.
* fix(computer-use): target Linux app windows reliably
Resolve app filters through the canonical cua-driver MCP app metadata and join running app PIDs back to windows. Preserve an exact selected window across capture_after, support direct capture by pid/window_id, and send the active window ID for coordinate pointer actions on Linux.
Co-authored-by: annguyenNous <annguyenNous@users.noreply.github.com>
Co-authored-by: grimmjoww578 <willies578@gmail.com>
Co-authored-by: ai-ag2026 <261867348+ai-ag2026@users.noreply.github.com>
* fix(computer-use): address review on PR #63725
Address three review comments from @f-trycua:
1. type_text, press_key, and hotkey now carry _active_window_id and
fail closed when it is missing. Previously they sent only the PID,
so CUA Driver fell back to the first window for that PID — input
could reach the wrong window in multi-window apps.
2. Coordinate scroll x/y are now capability-gated behind
input.scroll.coordinates. CUA Driver 0.7.1 Linux schema rejects
x/y on scroll; omitting them when the driver doesn't advertise
support avoids the schema rejection while still routing via
window_id.
3. Windows are sorted by z_index descending (higher = front, per CUA
Driver semantics) instead of ascending. Null z_index (Wayland) is
coerced to 0 in _ingest_windows so it doesn't crash the sort and
sorts to the back instead of being selected as the capture target.
---------
Co-authored-by: LeonSGP43 <cine.dreamer.one@gmail.com>
Co-authored-by: annguyenNous <annguyenNous@users.noreply.github.com>
Co-authored-by: grimmjoww578 <willies578@gmail.com>
Co-authored-by: ai-ag2026 <261867348+ai-ag2026@users.noreply.github.com>
The cross-turn stale-call circuit breaker (_check_stale_giveup in
agent/chat_completion_helpers.py) raises a RuntimeError when the
provider has been unresponsive for N consecutive stale attempts.
This error was classified as FailoverReason.unknown (retryable=True,
should_fallback=False), causing the retry loop to burn all max_retries
against the same dead provider — each retry hitting the circuit breaker
instantly with zero network overhead — before fallback was attempted.
Add a classification rule (section 7b in classify_api_error) that
recognizes the stale-breaker RuntimeError by its signature phrases and
classifies it as FailoverReason.timeout with retryable=False,
should_fallback=True. This makes the retry loop skip retries and go
straight to fallback provider activation on the first hit.
Test: test_stale_breaker_runtime_error_triggers_fallback_not_retry
The #54878 self-heal (SessionStore.get_or_create_session) drops a routing
key pointing at a session already ended in state.db and recovers/recreates
a fresh session_id under the same session_key. The #54947 fix (agent-cache
cache-hit guard in gateway/run.py) treats a cached agent whose snapshot
session_id differs from the current session_id, under the same
session_key, as an intentional /resume-/branch-style switch between two
live sibling conversations, and reuses it unchanged to protect the prompt
cache.
These two fixes compose incorrectly: when the #54878 self-heal just fired,
the cached agent's session_id is not a live sibling — it's the dead session
just routed away from. #54947's "different session_id -> reuse freely" rule
reuses it anyway. The stale agent runs the turn, and the post-run "session
split" sync (agent.session_id != session_id) then writes the routing key
straight back onto the dead session_id, undoing the self-heal. This repeats
on every subsequent message until an interrupt (e.g. /stop) happens to race
in before that post-run sync, silently discarding conversation context.
Reproduced live on the engineering gateway (2026-07-12, routing key
agent:main:telegram:dm:170829464:544520): 5 consecutive self-heal log lines
over ~40 minutes, each followed by the dead session_id being reused and
re-synced back, until an interrupted /stop finally let a fresh session
stick — at which point all prior context was gone.
No open upstream issue tracks this specific interaction as of 2026-07-12
(checked #54878, #54947, #59580, #59597, #61220 — all cover adjacent but
distinct edges of the self-heal / agent-cache system).
Fix: before applying #54947's reuse-on-mismatch rule, check (outside the
cache lock, via SessionStore._is_session_ended_in_db) whether the cached
snapshot's session_id is itself ended in state.db. If so, treat it as a
stale self-heal artifact and evict/rebuild fresh -- same as a genuine
cross-process write -- instead of reusing it. Re-validates the peeked
verdict against the tuple actually held under the lock so a race can't
apply a stale verdict to a different (possibly live) cache entry.
Tests: tests/gateway/test_stale_self_heal_agent_cache_eviction.py (5 new
cases: dead-session eviction, live-sibling reuse preserved [#54947 intact],
cross-process invalidation preserved [#45966 intact], same-session_id dead
edge case, lock-race re-validation). Full tests/gateway/ suite: 14 failed,
9040 passed, 11 skipped -- all 14 failures verified pre-existing on
unpatched main (confirmed via git stash + re-run), unrelated to this
change.
Reasoning ("Thinking") rendered through MarkdownTextContent with a smooth
typewriter reveal. TextMessagePartProvider mints a fresh part object on every
text change, and useSmooth resets its reveal to empty whenever the part
identity changes — so each reasoning delta restarted the animation from the
first character (h / hell / hello wo ...). Token-streaming reasoners
(R1/Qwen/GLM/Claude thinking) fire a delta per token and flash hard; GPT-5's
coarse reasoning summary updates too rarely to notice, which is why it looked
fine.
Drop smooth on the reasoning path so it plain-appends, exactly like the
assistant answer already does. Removes the now-dead smooth plumbing too.
GatewayRunner._prepare_inbound_message_text() interpolated
source.user_name — the platform-supplied, user-settable display name —
directly into the message text of every turn in a shared multi-user
session: f"[{source.user_name}] {message_text}". Shared sessions are the
default for any threaded conversation (thread_sessions_per_user defaults
to False) and apply to any group with group_sessions_per_user=False, so
no special configuration is needed to reach this path.
An unescaped display name containing embedded newlines could therefore
masquerade as a new markdown section (a fake "## Override" heading)
inside the live conversation the model reads on every turn — the same
indirect-prompt-injection vector gateway/session.py's
build_session_context_prompt() already guards against for the identical
user_name field via _format_untrusted_prompt_value(), which was never
applied to this sibling call site.
Add neutralize_untrusted_inline_text() alongside the existing helper in
gateway/session.py: it collapses embedded newlines/control characters to
a single inert line without JSON-quoting, so inline "[Name] message"
formatting is preserved byte-for-byte for the common case (unlike
reusing _format_untrusted_prompt_value directly, which would add visible
quote marks to every sender prefix). Wire it into the sender-prefix
construction in gateway/run.py.
Generalize the clarify opt-out into an UNBOUNDABLE_TOOLS set so a run
containing image_generate also stays a plain, fully-visible stack instead
of collapsing into the bounded window, where the max-height + gradient mask
clipped the image the same way it clipped clarify forms.
When an agent turn finishes while the user is viewing a different
session, the sidebar now shows a steady green dot on that session —
distinct from the blue pulsing dot of a running turn and the gray dot
of an idle one. Opening the session clears the indicator.
The unread state is ephemeral renderer-side state, matching the
existing $workingSessionIds and $attentionSessionIds pattern: no
persistence, no backend involvement, wiped on gateway-mode switch.
Co-authored-by: liuhao1024 <liuhao1024@users.noreply.github.com>
Co-authored-by: dschnurbusch <dschnurbusch@users.noreply.github.com>
Co-authored-by: Flow Digital Inc. <flow-digital-ny@users.noreply.github.com>
Bug 2 of #28156: the picker offered us./global. inference profiles to
EU-region endpoints (unroutable — AWS rejects them regardless of
credentials) and _RECOMMENDED hardcoded us.anthropic.* ids, so non-US
pickers pinned profiles their endpoint can't invoke.
- bedrock_model_routable_from_region(): geo-prefixed profiles are only
offered in their own geography (full AWS prefix set incl. apac./jp./
ca./sa./me./af.); bare ids and global.* pass everywhere; unknown
region shapes hide nothing.
- Recommendations match geo-agnostically on the base model id, so an EU
picker pins eu.anthropic.claude-sonnet-4-6; in-region geo profiles
sort above global.* for the same model (addresses the global.*-first
ordering complaint from the issue thread).
- Dedup generalized from (us., global.) to all profile prefixes.
Rewrites the cherry-picked test to import os locally and monkeypatch the
resolver seams instead of patching bedrock_adapter internals; adds the
inverse assertion (no bearer -> AnthropicBedrock SDK path preserved).
Three fixes for the Bedrock Claude path:
1. Streaming fallback: When AnthropicBedrock SDK raises 'Unexpected event
order' (SDK misparses Bedrock error events as message_start), auto-switch
to native Converse API for the rest of the session instead of failing
after 3 retries.
2. Image base64 decode (#33317): data URL payloads were passed as base64
strings to source.bytes, but boto3 re-encodes at the wire layer. Now
decoded to raw bytes before passing to Converse API.
3. Bearer token routing (#28156): Users with AWS_BEARER_TOKEN_BEDROCK are
now routed through Converse API regardless of model, since the
AnthropicBedrock SDK only supports SigV4 signing.
3 new tests. 121 bedrock_adapter tests passing.
The docs search theme (@easyops-cn/docusaurus-search-local) defaults
fuzzyMatchingDistance to 1, so every term also matched words one edit
away. Two user-visible failures on the 14.4 MB production index:
- Wrong results: 'keet' returned 'Microsoft Teams Meetings',
'google_meet', 'Keep the Model Loaded' etc. — 'meet' and 'keep' are
both one edit from 'keet', and the stemmer indexes 'meetings' as
'meet'.
- Search appearing to die: fuzzy matching multiplies the generated
lunr queries (distance matrix x maybe-typing variants x
leave-one-out terms — up to 210 queries per keystroke on multi-word
input), and fuzzy REQUIRED terms are the expensive scan kind. A
typo'd 3-word query stalled the single-threaded search Web Worker
for 25+ seconds; every later keystroke's search queued behind it,
so the bar stopped returning results.
Setting fuzzyMatchingDistance: 0 keeps exact-word-or-prefix semantics
(keet -> keet*), which is the behavior users asked for. Validated by
running the plugin's shipped smartQueries/tokenize code against the
downloaded production search-index.json: legitimate queries (cron,
telegram, prefix 'memor') return identical results; worst-case typo
queries drop from 210 queries / 365-532ms per keystroke to 50-105 /
53-188ms; false 'keet' matches gone.
Follow-up hardening on top of #64158 (@DavidMetcalfe):
Backend (the root-cause fix):
- hermes_cli/moa_config.py: add validate_moa_payload() — strict write-time
counterpart to the deliberately tolerant normalize_moa_config(). Flags
half-filled slots, empty reference lists, recursive moa slots, naming the
exact preset/slot.
- hermes_cli/web_server.py: PUT /api/model/moa validates before normalizing
and returns 422 with the specific problems instead of silently swapping the
user's preset for hardcoded defaults (#64156). Also declares
fanout / reference_max_tokens / reasoning_effort on the Pydantic payload so
client round-trips no longer erase hand-set values.
Desktop:
- Replace sanitize-then-send with hold-while-incomplete: the debounced
autosave is deferred (not repaired) while any slot is half-filled, and
flushes once the model pick completes the edit. Mid-edit UI state is never
repainted by a save response (generation guard covers held edits too).
- updateMoaSlot only clears the model when the provider actually changed.
- Explicit preset ops (set default / add / delete) cancel the pending
autosave and invalidate in-flight responses so the two writers can't race.
- Stable row keys (preset+index) so mid-edit rows don't remount; cleared
model shows the 'Model' placeholder instead of vanishing.
Both TS clients' MoaConfigResponse types now declare the round-tripped
fields (fanout, reference_max_tokens, reasoning_effort).
Tests: 12 new backend unit tests (validate_moa_payload contract incl.
validate/normalize agreement), 3 new web_server endpoint tests (422 on
half-filled ref/aggregator, fanout round-trip), 3 new desktop vitest cases
(autosave held while half-filled, flush on completion, same-provider
reselect no-op). E2E validated against a live TestClient with isolated
HERMES_HOME: bug sequence now 422s with config untouched.
Fixes#64156
Three fixes in model-settings.tsx:
1. Filter incomplete slots before autosave: sanitizeMoaRefsForSave() strips
reference slots with empty model before any autosave (the 600ms debounce
was sending half-filled provider-but-no-model slots to the backend, where
_clean_slot rejects them and _normalize_preset falls back to hardcoded
defaults). Presets with zero valid refs keep their empty reference_models
array rather than being silently dropped. The aggregator slot is also
sanitized when its model is empty.
2. Add withActive() to MoA provider dropdowns: the reference and aggregator
provider Selects filtered to authenticated-only, so unauthenticated
current values (e.g. openai-codex) rendered blank. Mirror the existing
pattern from the model Select.
3. Add generation counter to scheduleMoaSave(): stale save responses could
overwrite newer state. Bump a counter on each save and skip setMoa/setError
if a newer save was scheduled in the meantime.
Follow-ups on the salvaged #20096 profile-routing feature:
- _profile_name_for_source now returns None unless gateway.multiplex_profiles
is on. Routing stamps source.profile, which namespaces session/batch keys,
but the profile-scoped agent run only activates under multiplexing — without
the gate, configured routes with multiplexing off split batch/session keys
into agent:<profile> while the agent still ran from agent:main.
- Widen the profile-aware _text_batch_key fix from Discord to every adapter
that builds batch keys via build_session_key (telegram, whatsapp, matrix,
feishu, wecom, weixin) — routing is platform-generic, so the batch-key
namespace fix must be too.
- Downgrade the no-route-matched log from INFO to DEBUG (fired on every
unrouted inbound message).
- GatewayConfig.to_dict(): serialize profile_routes as plain dicts
(ProfileRoute dataclasses are not JSON-safe).
- Docs: correct the 'independent of multiplexing' claim in
docs/profile-routing.md (routing requires multiplexing), fix the
platform-only specificity row (0, not 1), and document profile_routes in
website/docs/user-guide/multi-profile-gateways.md.
- Tests: pin the multiplex gate (routes ignored when off, active when on,
build_source end-to-end stays in agent:main when off).
Completes the review's ask for "adapter-to-session-key integration coverage
for Discord and a non-Discord platform" on #20096.
Drives a concrete adapter's real BasePlatformAdapter.build_source with an
injected gateway_runner, asserts the matched route's profile is stamped on
the source, and that build_session_key scopes the key under agent:<profile>:
(versus the shared agent:main: namespace). Covers Discord and Telegram — the
Telegram case is the bug-#2 path that previously fell through to default.
Adds a regression anchor: without gateway_runner, profile stays None and the
key lands in agent:main (the silent fallback the fix removes for non-Discord).
Co-Authored-By: Claude <noreply@anthropic.com>
Addresses hermes-sweeper review on #20096.
Problem 1 (profile_routing.py): route matching returned True on a chat_id
hit before the guild_id constraint was consulted, so a route declaring both
guild_id and chat_id matched on chat_id alone. Restored conjunctive (AND)
semantics — every declared discriminator must hold; hierarchical parent_chat_id
matching is preserved. Added a regression test for the guild+chat case.
Problem 2 (base.py / run.py): gateway_runner was injected only when an adapter
pre-declared the attribute, and only Discord did — so build_source never called
_profile_name_for_source for Telegram/Feishu/Slack/etc., despite the
platform-generic claim. Declared gateway_runner on BasePlatformAdapter and made
the plugin-registry injection unconditional, so profile routing now reaches
every platform. Added non-Discord (Telegram) resolution coverage and an
injection-inheritance test.
Also adds docs/profile-routing.md documenting gateway.profile_routes
(matching rules, specificity, profile isolation) — requested in review.
Co-Authored-By: Claude <noreply@anthropic.com>
Two follow-ups observed after deploying profile routing:
1. sessions.profile_name was NULL even when the agent ran inside the
routed profile scope. _insert_session_row never wrote it,
get_or_create_session / reset_session never passed it through, and
the agent-side _ensure_db_session fallback had no way to read it.
- Declare profile_name TEXT in SCHEMA_SQL so _reconcile_columns
auto-adds it on existing DBs.
- _insert_session_row takes profile_name and writes it.
- SessionStore passes source.profile (or old_entry.origin.profile
on reset) into db_create_kwargs.
- _ensure_db_session reads the active profile via
get_active_profile_name() inside _profile_runtime_scope.
2. DiscordAdapter._text_batch_key called build_session_key without
profile=, so the batch key always landed in agent:main even when
the routed profile differed — diverging from the agent session
key namespace (agent:crypto-trader, agent:ai-expert, ...).
Pass event.source.profile through so both namespaces agree.
Live verification (jth-server-2, 2026-06-28): a test message in a
routed #coin thread produced agent:crypto-trader:discord🧵...
in the batch log and profile_name=crypto-trader in the sessions
row. Default-routed chat still produced agent:main / NULL.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
load_gateway_config only forwarded the top-level multiplex_profiles
key, ignoring the nested gateway.multiplex_profiles form. The latter
is what `hermes config set gateway.multiplex_profiles true` writes,
so users who ran that command got multiplex_profiles=False silently
— no warning, no fallback, profile_routes just stopped matching.
Loader now checks the top-level key first, falls back to the nested
gateway section, and only then defaults to False. Same precedence is
applied to other nested-form keys (profile_routes already did this).
Tests cover: top-level honored, nested honored (regression test for
the silent-fallback bug), default False, top-level overrides nested.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
_adapter_credential_fingerprint only looked at adapter.token directly,
but Discord (and similar) adapters store the bot token on their config
sub-object, not on self. Every Discord adapter in a multiplexed
gateway therefore returned None, the same-token conflict check was
silently skipped, and N adapters all polled the same bot token —
producing a per-message race where whichever adapter won the GIL
answered the user.
Adds a config-token fallback (token, then bot_token) so the check
actually fires for config-backed adapters. Direct adapter.token
still takes precedence when both exist.
Tests cover: config-backed token produces a fingerprint, distinct
tokens produce distinct fingerprints, direct token wins over config,
config without token attributes returns None.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
STANDARD_PROFILES, normalize_profile, validate_profile_name, and
is_standard_profile in hermes_constants were superseded by
hermes_cli.profiles.{normalize_profile_name, validate_profile_name}
but never removed. profile_routing.py is updated to import from the
canonical location; the old helpers are deleted.
Lazy import inside parse_profile_routes avoids the circular dependency
at module load time (hermes_constants -> hermes_cli -> hermes_constants).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Three follow-ups to the initial routing PR after code review:
1. Remove dead forum-post hierarchy cache. The `_forum_post_cache`,
`register_forum_post()`, and `resolve_forum_channel()` were never
wired up — no caller in the codebase. Discord's adapter already
sets `parent_chat_id` to the immediate parent (forum channel for a
forum post), so the existing `self.chat_id == parent_chat_id`
branch in `matches()` handles forum posts correctly without a
cache. The hierarchical-resolution branch in `matches()` and the
bounded-LRU infrastructure are removed.
2. Fix docstring specificity numbers (8 → 14, 4 → 6) and rewrite the
"Hierarchical matching" section to describe the actual one-level
parent_chat_id behavior. Removed unused `OrderedDict` and `Set`
imports.
3. Warn loudly when a routed profile doesn't exist on disk. Previously,
a typo in `profile_routes` (e.g. `crypto-tradr`) silently fell back
to the global HERMES_HOME, causing the message to read the default
profile's memory/credentials with no signal to the operator. Now
emits a `logger.warning` with the profile name, source identifier,
and the fallback reason. Bare-exception path also gets `exc_info`.
Tests:
- test_profile_routing.py: +2 tests verifying forum post matching via
direct parent_chat_id (covers the case the removed cache was meant
for). 31 total, all pass.
- test_profile_resolution.py: NEW, 12 tests covering resolution order
(source.profile > routing > active > default), missing-profile
warning, exception handling, and routing consultation. All pass.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Adds gateway.profile_routes config that routes specific Discord
guilds/channels/threads (and other platforms) to different profiles.
The routing engine uses hierarchical specificity matching
(thread > channel > guild) with bounded LRU caching for forum post
resolution.
Routing result is stamped on source.profile by BasePlatformAdapter
.build_source() at inbound time. When gateway.multiplex_profiles is on,
the existing _profile_runtime_scope machinery picks up source.profile
and runs the whole turn inside the profile's HERMES_HOME — so memory,
skills, config, and secrets all resolve to that profile automatically.
No new isolation code is added; this PR only adds the routing decision
layer on top of the existing multiplexing infrastructure.
Configuration:
gateway:
multiplex_profiles: true
profile_routes:
- name: server-default
platform: discord
guild_id: "GUILD_ID"
profile: server-profile
- name: special-channel
platform: discord
guild_id: "GUILD_ID"
chat_id: "CHANNEL_ID"
profile: channel-profile
When multiplex_profiles is off, profile_routes is ignored (no behavior
change for single-profile gateways).
Tests: 29 unit tests covering specificity scoring, hierarchical
matching, path-traversal validation, and config parsing.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The optional blender-mcp skill predates the blender MCP catalog entry
(#64463) and taught the agent to hand-roll raw TCP JSON to the addon's
socket on port 9876 from execute_code — bypassing the catalog's version
pinning and install-time tool curation.
Reworked to v2.0.0 as the companion skill for the catalog entry:
- prerequisites now go through 'hermes mcp install blender'
- interaction surface is the four curated MCP tools, not a raw socket
- keeps the valuable content: addon setup, bpy recipes (materials,
keyframes, render-to-file), pitfalls (timeouts, absolute paths,
object mode), plus new pitfalls (xvfb headless, no-sandbox warning,
remote-host path resolution)
- explicit anti-pattern note: do not hand-roll TCP to 9876
- description shortened to <=60 chars per skill authoring standards
alireza78a's original bpy patterns and pitfalls are preserved and
credited. Docs page regenerated via generate-skill-docs.py (scoped to
this skill only; unrelated generator drift left untouched).
Two consecutive incomplete assistant interims with identical visible
content (content + reasoning) are collapsed even when opaque provider
state (encrypted reasoning item ids, message item phases) drifts per
continuation — previously that drift defeated dedup and caused message
storms (#52711). The latest opaque payload is written onto the existing
message in place, so continuation replay still uses fresh provider
state.
Salvage note: the original PR also made the 3-retry continuation cap
cumulative per turn (no reset on progress). That half is intentionally
dropped — a legitimate long turn alternating incomplete/progress would
hard-fail at 3 cumulative, changing semantics beyond the reported bug.
Follow-up to the salvaged #51657: the conversation loop returns the
retry-exhaustion sentinel as BOTH final_response and error, so the
original detector (which required final_response to be falsy) never
fired on real exhaustion turns — the sentinel text was delivered
verbatim into the channel, exactly the #51628 poisoning vector. Detect
the sentinel echo, blank it before empty-response normalization, and
never suppress a turn whose final_response is genuine model text.
Also: dedupe-guard mock fix in the test fixture (has_platform_message_id
must return False, not a truthy MagicMock) and two guard tests
(real answer never suppressed; interrupted/failed never classified).
Map Codex Responses status=incomplete with incomplete_details.reason=content_filter to finish_reason=content_filter so the existing refusal/fallback path runs instead of burning incomplete continuation attempts.
Follow-up on @AlexFucuson9's cherry-picked multimodal flattening:
- http(s) image parts render as '[image: <url>]' so the summary keeps a
referenceable handle after compaction (base64 data: URLs still
collapse to '[image]' — no reusable reference, and leaking them is
the bug being fixed)
- unknown part types (document, future shapes) render as '[<type>]'
instead of being silently dropped
- test docstrings corrected: the original crash premise is stale
(redact_sensitive_text grew str() coercion in #52147); the live bug
is repr-noise + base64 leakage into the summarizer input
msg.get('content') can return a list of parts for multimodal messages
(containing text, images, etc.). The old code passed this list directly
to redact_sensitive_text(text: str), which raised AttributeError on
list.replace(), causing context compression to fail entirely for any
session with attached images.
Fix: detect list content and extract text parts before redacting.
Image parts are replaced with '[image]' placeholder.
ReviewDirRow never rendered a DiffCount, so collapsed folders gave no
indication of the additions/deletions inside them. The tree builder
already aggregates added/removed onto directory nodes (verified by
tree-data tests) — this just renders them, matching the file rows.
CI runners have no global git identity. The remote repo seed commit
needs inline -c user.email/user.name flags, matching the pattern
already used by ensureGitRepo.
When "new worktree" branches off a remote-tracking ref like origin/main,
`git worktree add -b <branch> <dir> origin/main` auto-sets upstream
tracking (branch → origin/main), producing `branch:origin/main` in branch
listings. The user wants a standalone local branch — like `git checkout
origin/main && git checkout -b branch` — not one silently wired to the
remote. Add `--no-track` when the base is an `origin/` ref. Local branch
bases are unaffected (they never triggered tracking).
Follow-up on the salvaged fallback: silent degradation is how #64333 went
unnoticed (jobs dead on arrival, only errors.log knew). Warn once with a
resync hint, and cover both the skewed and healthy paths with tests.
Extract shared hermes_cli/input_sanitize.py (bracketed-paste wrapper stripping
and terminal ~[[e artifact suffix collapse) and wire it into prompt.submit
and the Desktop composer so corrupted user text is cleaned before
messages.content is persisted.
* fix(auxiliary): route direct-create aux callers through call_llm (#35566)
Five callers (kanban_decompose, kanban_specify, profile_describer, and
goals.py's judge + draft-contract) built raw clients via
get_text_auxiliary_client() and passed extra_body=get_auxiliary_extra_body()
— which only returns Nous portal tags and ignores
auxiliary.<task>.extra_body from config.yaml entirely. That was the
remaining half of #35566 after the call_llm path was fixed.
Routing them through call_llm(task=...) gives each caller the full
auxiliary contract for free: task extra_body, the reasoning_effort
shorthand, transient retries, provider-profile projection, and fallback
chains. goal_judge gains a DEFAULT_CONFIG block (it had none — its
provider/model overrides silently didn't exist as documented keys).
get_auxiliary_extra_body() now has zero non-test callers; kept for
plugin back-compat.
Fixes#35566.
* test: migrate kanban dashboard + CLI specify mocks to call_llm
Two more consumers of specify_task mocked the old
get_text_auxiliary_client symbol (missed in the first sibling sweep —
they live outside tests/hermes_cli's kanban files): the dashboard
plugin's /specify endpoint tests and the /kanban slash-command E2E.
Same migration as the rest: mock call_llm at the source, no-provider
now surfaces via the LLM-error branch.
Ollama's /v1/chat/completions silently ignores extra_body.think (it only
honours it on /api/chat — ollama/ollama#14820), so agent.reasoning_effort:
none never actually disabled thinking on OpenAI-compatible Ollama routes.
Emit the top-level reasoning_effort='none' field (which Ollama respects)
alongside think=False (kept for proxies and the native /api/chat path).
The PR's second half (propagating reasoning_config to the background-review
fork) already landed on main via agent/background_review.py, so only the
provider-profile change is salvaged here, resolved onto the current
GLM/effort-aware profile.
Salvaged from PR #29820 by @Epoxidex.
Five tests for the salvaged #37217 Bug B fix: vendor-field passthrough,
reasoning-key + private-key exclusion, merge-over-existing (fast-mode
speed), no-extra_body regression guard, reasoning-only adds nothing.
Live probes against api.anthropic.com informed the exclusion design:
Anthropic strictly validates the request body (unknown keys 400 with
'Extra inputs are not permitted'), so the passthrough forwards only
caller-configured fields and never the OpenAI-shaped reasoning dict
(translated natively) or _-private plumbing keys.
Two related bugs in _AnthropicCompletionsAdapter.create() in
agent/auxiliary_client.py silently discard caller-supplied
reasoning_config and extra_body on the Anthropic-Messages
auxiliary-protocol path:
* Bug A: reasoning_config=None was hardcoded at L1000, so the
reasoning_config parameter on build_anthropic_kwargs was
unreachable for any auxiliary task. The main agent path
(agent/transports/anthropic.py) already reads
reasoning_config from caller params; this PR aligns the
auxiliary adapter with the same pattern.
* Bug B: create(**kwargs) accepts an OpenAI-style kwargs
payload from the caller but only forwards a hand-picked
subset to self._client.messages.create(). Any caller-supplied
extra_body (e.g. thinking control, metadata, service_tier,
vendor-specific fields) was dropped on the floor. The
codex/responses transport in the same file already merges
extra_body; the Anthropic branch is the gap.
This unlocks the caller-supplied extra_body path so auxiliary
callers can set per-vendor request fields (including
thinking: {type: "disabled"} for Anthropic-compatible vendors
that require an explicit disable on the wire), and lets the
reasoning_config kwarg flow into build_anthropic_kwargs like the
main agent does. Both changes are backward-compatible for
callers that don't pass the affected kwargs.
Affected providers (all routed through _AnthropicCompletionsAdapter
via _maybe_wrap_anthropic): anthropic (native), minimax /
minimax-cn, kimi-coding / kimi-coding-cn, z.ai / GLM, and any
custom /anthropic-suffixed endpoint. See PR description for
related issues (#35566, #7209, #16533, #32813, #29248).
Catalog entries now follow the same supply-chain rules as pyproject
dependencies:
- n8n: install.ref main -> full commit SHA 7a9ae007 (2026-05-23,
branches/tags can be moved by the upstream owner; SHAs cannot)
- new contract test: every shipped manifest must pin exactly —
git installs need a 40-char SHA, uvx/npx-style launchers need
pkg==X / pkg@X with a digit-leading version (rejects bare names,
ranges, and npm dist-tags like @latest)
- module docstring documents the pin policy (exact version, 2-week
cooldown)
unreal-engine and linear are http transports (server runs elsewhere)
so there is nothing to pin at the transport layer.
Verified: unpinning blender-mcp in the manifest makes the contract
test fail with a named diagnostic; restoring the pin passes.
MCP catalog entries follow the same supply-chain rules as pyproject
dependencies: exact version pin, and the pinned release must be at
least 2 weeks old. blender-mcp 1.6.4 released 2026-06-11 (~5 weeks
old, also the latest release). uvx now resolves the exact version
instead of latest-at-launch.
Adds optional-mcps/blender (ahujasid/blender-mcp, stdio via uvx). The
server advertises 22 tools; 18 front optional asset services with no
upstream trim mechanism, so tools.default_enabled pins the install to
the core surface (scene/object info, viewport screenshot, code exec)
and the rest stay opt-in through 'hermes mcp configure blender'.
Manifests can now declare transport.env (static, non-secret subprocess
env vars), parsed/validated in _parse_manifest and written by
_build_server_config — used here to ship DISABLE_TELEMETRY=true per
the no-telemetry-without-opt-in policy. Runtime already honored
per-server env; manifests just couldn't declare it.
With the TTFB watchdog now scaled (not disabled) for large requests on
main, the hard ceiling is a backstop against mid-stream wedges, not the
primary stall detector. The original 600s default clamped BELOW the
intentional 1200s stale floor for >100k-token requests, partially
reverting the floor that keeps healthy gateway-scale payloads alive.
Raise the default to 1500s so the ceiling only catches unbounded
growth, never healthy slow turns.
A large Codex request (estimated context >= 10k tokens) disables the no-byte
TTFB watchdog on purpose, and openai_codex_stale_timeout_floor *raises* the
stale timeout (up to 1200s at >100k tokens) so healthy gateway-scale payloads
aren't aborted mid-prefill. When the backend genuinely stalls — no first byte
AND no events, exactly the #64507 symptom — the request is only reclaimed at
that high stale floor, so the session can hang for 13+ minutes with an idle
slash_worker and no ended_at/end_reason while Desktop still shows it as active.
Add a flat, finite hard ceiling on total openai-codex request time that always
applies (min() of the computed stale timeout and the ceiling) regardless of the
TTFB-disable / stale-floor interaction. A stalled large request is now killed
at the ceiling and the retry loop / visible failure path takes over instead of
hanging indefinitely. Tunable via HERMES_CODEX_HARD_TIMEOUT_SECONDS (default
600s; 0 disables the ceiling to restore pre-fix behavior).
Closes#64507
TRUNCATE checkpoint every 50 writes causes B-tree corruption on large
databases (65K+ pages) due to the exclusive-lock I/O pressure from
checkpointing thousands of frames at once.
Switch periodic _try_wal_checkpoint() to PASSIVE mode which does not
require an exclusive lock and cannot corrupt pages under I/O pressure.
Keep TRUNCATE in close() and pre-VACUUM paths where it is safe
(infrequent, controlled conditions).
Also replace silent `except Exception: pass` with logged warnings for
checkpoint failures so operators can detect early corruption signals.
Fixes#45383
_summarize_tool_result() built the web_extract summary straight from the
first `urls` entry. When web_search results are forwarded into web_extract
(a common chain), that entry is a dict ({"url"/"href": ...}) rather than a
URL string. With 2+ URLs the `url_desc += " (+N more)"` step then raised
`TypeError: unsupported operand type(s) for +=: 'dict' and 'str'`, aborting
the pre-compression pruning pass; with a single URL the raw dict repr
leaked into the summary text.
Unwrap the URL from dict entries before use — mirroring the existing
web_extract handling in agent/display.py (`_display_url`),
tools/web_tools.py (`_web_extract_url`) and acp_adapter/tools.py
(`build_tool_title`) — so `url_desc` is always a string and the
concatenation is always str + str.
Adds regression tests covering multiple/single dict URLs, the href key,
malformed dicts, and the plain-string path.
test_normalize_codex_response_salvage_is_xai_scoped broke on main when
two same-day merges crossed: #64764 (#64434 — trust response.status for
reasoning-only turns on UNRECOGNIZED Responses backends) changed what a
bare _normalize_codex_response(response) call returns for
status='completed' reasoning-only output (now 'stop'), while #64768
added this test calling with no issuer_kind and expecting 'incomplete'.
The test's intent is that the xAI reasoning-channel salvage does not
leak into other special-cased backends — pin issuer_kind='codex_backend'
so it exercises exactly that (same pattern as
test_normalize_codex_response_treats_summary_only_reasoning_as_incomplete,
which was already pinned for #64434).
Add merge_existing to save_config (default False for full-document callers
like the dashboard YAML editor) and route partial writes through
_merge_partial_save. _persist_migration writes the full migrated dict
directly so deleted keys are not resurrected from the on-disk file.
resetLayoutTree() claims to "restore everything" but never reopened collapsed
SIDES. A sidebar hidden before a reset survived it, so the next ⌘B toggled from
that stale-hidden state into a SHOW — and the user's hide never persisted across
reload. Reopen every bound side through its store so the toggles stay truthful.
Adds a store-level regression test (real modules, re-import = reload) covering
hide→reload and the reset→hide→reload repro.
Drive the composer's layout off its OWN measured width (the existing
ResizeObserver, so it reacts per-tile, not per-viewport) instead of only
stacking or overflowing:
- >=440px: full inline row, full model label.
- 320-440px: still inline, but the model pill sheds its label for its chevron
icon (~120px back) so the controls stop crowding the placeholder.
- <320px: stacks to the two-row layout, pill stays iconized.
Adds COMPOSER_COMPACT_PILL_PX above the stack breakpoint and a `compactPill`
signal from useComposerMetrics, wired to the model pill's existing compact mode.
Community reports of GPT 'infinitely thinking' are usually a slow or
overloaded provider plus silent retry machinery: the CLI/TUI/Desktop
spinner shows a generic 'cogitating...' verb for the whole wait and the
gateway heartbeat says only 'Working — N min'.
Add AIAgent._emit_wait_notice(): rewrites the live spinner/status line
(thinking_callback → CLI prompt_toolkit widget, thinking.delta → TUI +
Desktop) and updates the activity tracker (included in the gateway's
'⏳ Working — N min' heartbeat). Wired at the four wait points:
- non-streaming wait loop: after 30s with no response, the line becomes
'⏳ waiting on <model> — Ns with no response yet (provider may be slow
or overloaded; auto-reconnect at Ns)'
- streaming wait loop: same explanation after 30s with no chunks,
including the long-thinking case
- TTFB / stale-stream kills: '⚠ no response from provider in Ns —
reconnecting...'
- Codex continuation retries: '↻ model returned reasoning with no final
answer — asking it to continue (n/3)' instead of silence
Notices are fail-open (display errors never break the wait loop) and
gateway sessions without a display callback still get the improved
activity description.
Make every session tab speak the same drag language as a sidebar row, and
keep tab focus 1:1 across the sidebar, tiles, and main:
- Drag a session tile's OR the main workspace tab onto a composer to link the
chat (@session chip), or onto a zone/edge to stack/split — tabDrag now
returns whether it took the drag, so the workspace tab defers to the generic
pane move on a fresh draft (nothing to link).
- A lone tool panel (terminal/logs) dragged to its own zone keeps its header,
so it stays draggable/closable instead of stranding a dead, tab-less zone.
- The sidebar highlight follows the FOCUSED session (interacted tile, else the
main selection) rather than only the main one.
- Clicking an already-open session jumps to its tab — an open tile, or the
workspace tab when it's the main session and focus sits on a tile — instead
of a dead no-op reload.
- Secondary (single-chat pop-out) windows boot to the default tree with no
tiles and never persist their stripped-down layout back, so they can't
inherit or clobber the primary window's tabs/splits.
The pointer drag ghost is extracted to a shared lib/drag-ghost and the drop
affordances drop their now-unused labels.
Follow-up to the salvaged #63690: when the interim assistant message is
too empty to append (no content and no reasoning of any kind), the last
message in history is still the prior user/tool turn — appending the
user-role nudge there would create a user→user or tool→user sequence
that strict providers reject. Only append the nudge when the last
message is an assistant turn. Also add IpastorSan to AUTHOR_MAP.
grok-4.x on the xAI /v1/responses surface sometimes ends a turn with only
reasoning items — no message output item, no tool calls — and those
reasoning items carry no encrypted_content. Two compounding problems:
1. The model occasionally emits its final answer INSIDE the reasoning
channel, delimited by grok's internal "<response>" tag. The answer
exists but is classified reasoning-only → finish_reason=incomplete.
2. An interim assistant message holding only plain-text reasoning replays
as nothing in _chat_messages_to_responses_input, so every continuation
request is byte-identical to the one that just failed. The model
deterministically repeats the reasoning-only response until the retry
budget is exhausted and the turn dies with "Codex response remained
incomplete after 3 continuation attempts".
Fixes:
- _normalize_codex_response (xai_responses only): salvage the
<response>-delimited tail from the reasoning text and promote it to
assistant content; the untagged prefix stays as thinking text.
- Codex-incomplete continuation path: when the interim message has
nothing the input converter will replay (no content, no encrypted
reasoning items, no message items), append a user-role nudge so the
retry actually differs and explicitly asks for the final answer /
pending tool call. Mirrors the existing _get_continuation_prompt
pattern used for length truncation.
Observed live with grok-4.20 on xai-oauth (2026-07-13); sibling of the
grok-composer web_search incomplete-loop fix in transports/codex.py.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The #64434 change makes unrecognized issuers trust
response.status='completed' for reasoning-only turns, so this sibling
test (which exercised the old default-path behavior) now pins the Codex
backend explicitly — the surface where reasoning-only still means
'still thinking'.
Follow-up to the salvaged #64449: the status-trusting branch flipped
github_responses to 'stop' alongside unknown relays. Copilot fronts the
same OpenAI model family as codex_backend and shows the same
reasoning-only 'still thinking' degeneration, so it stays on the
continuation path. Only unrecognized (other:*) backends trust
response.status='completed' as terminal.
The aggregator is MoA's acting model, but the main loop's reasoning
gates key off the virtual moa://local identity and never fire — so with
no per-slot reasoning_effort the aggregator silently ran at the backend
default, ignoring the user's reasoning config entirely (#64187).
New _aggregator_reasoning_config(): slot value > full acting-model
resolution via the shared chokepoint (agent.reasoning_overrides for the
slot's model > global agent.reasoning_effort; YAML False stays
'disabled'). Applied to both aggregator call sites (acting turn +
one-shot /moa synthesis).
Reference advisors intentionally keep slot-or-default: inheriting a
global xhigh into every advisor fan-out would silently multiply cost.
Fixes#64187.
The remote model catalog (website/static/api/model-catalog.json) now labels
exactly one entry per provider block with "default": true — z-ai/glm-5.2 for
both OpenRouter and Nous Portal. That labeled entry is the model Hermes
silently lands on when the user never picked one, and it can be rotated by
editing the manifest alone: no release needed.
- model_catalog.py: get_default_model_from_cache() reads the label from the
in-process/disk cache only — never triggers a network fetch, so hot
resolution paths (agent build, gateway session setup) stay network-free.
- models.py: get_preferred_silent_default_model() resolves catalog label
first, PREFERRED_SILENT_DEFAULT_MODEL constant second (offline/fresh
install). _PROVIDER_SILENT_DEFAULT_OVERRIDES dict replaced by
_SILENT_DEFAULT_PROVIDERS routing through the shared resolver.
fetch_openrouter_models() preserves the "default" badge through live
/v1/models refreshes so the picker shows it.
- scripts/build_model_catalog.py: generator emits the default label so
regeneration can't drop it.
- website/docs/reference/model-catalog.md: schema documents the new field.
- Salvaged from PR #61141 (@HumphreySun98): bare-provider /model switches
(/model nous) route through the cost-safe default instead of curated
entry [0].
- tests: catalog-label precedence, constant fallback, stale-label fallback,
cache-only (no network) guarantee, and a shipped-manifest contract test
pinning the labeled entry to PREFERRED_SILENT_DEFAULT_MODEL.
E2E (temp HERMES_HOME): fresh-install constant fallback, shipped-manifest
label read, release-free rotation (relabeled cache -> new default across
models.py, tui_gateway, and gateway empty-model paths) all verified.
`detect_static_provider_for_model` handles a bare provider name typed as a
model (e.g. `/model nous`) by returning `_PROVIDER_MODELS[provider][0]` — the
first curated entry. For metered aggregators whose curated list is ordered
most-capable-first (Nous Portal), entry [0] is the priciest flagship, so
`/model nous` silently switched to it.
This is exactly the billing footgun `_PROVIDER_SILENT_DEFAULT_OVERRIDES` /
`get_default_model_for_provider` exist to prevent (per their docstring, a
missing model "escalated to Opus and billed 863 requests before the user
noticed"). The non-interactive fallback already routes through that cost-safe
helper; the interactive `/model <provider>` path did not.
Route this path through `get_default_model_for_provider` too. Providers
without a silent-default override are unchanged (the helper returns
`models[0]`), so only overridden providers (currently `nous`) change — from
the flagship to the low-cost default.
Adds regression tests: `/model nous` resolves to the cost-safe default, and
non-overridden providers still resolve to their first catalog model.
Consolidation follow-up on top of @liuwei666888's compressor fix (#52431)
and @Frowtek's ACP render guard (#62588) — completes the class fix across
every display-layer consumer of tool-call arguments:
- agent/context_compressor.py: wrap _summarize_tool_result in a
never-raises backstop (same pattern as get_cute_tool_message and the
ACP guard). The per-branch _str_arg coercions keep summaries
informative; the wrapper guarantees compression can never crash-loop
on a summary branch we didn't anticipate. Also reject non-dict parsed
args (a JSON list/scalar would crash every args.get call).
- agent/display.py: build_tool_preview's process branch sliced
session_id/data without coercion — crashed build_tool_preview and
build_tool_label on the live tool-progress callback (probed live:
TypeError 'int' object is not subscriptable). Coerce like the
sibling branches.
- tests: backstop fuzz matrix (16 tools x 6 hostile value shapes),
fallback-shape contracts, display process-preview regressions,
non-dict args guard.
build_tool_start renders every ACP (Zed) tool call — on the live tool-progress
callback (acp_adapter/events.py) and during session history replay
(acp_adapter/server.py). It called build_tool_title and extract_locations
directly, so a model that emits a malformed argument crashed the render:
- terminal `command` as null/number -> TypeError (len() in build_tool_title)
- delegate_task `goal` as a number -> TypeError (len())
- read_file `path` as a non-string -> pydantic ValidationError building a
ToolCallLocation
A live crash breaks the tool-call event; a persisted one breaks history replay
on every resume of that session. The sibling CLI label builder
get_cute_tool_message was already wrapped for exactly this reason
(agent/display.py: "display must never abort a turn").
Wrap build_tool_start the same way: on any builder failure, fall back to a
minimal, valid start event (tool name as title, resolved kind). The happy path
is unchanged.
Adds tests for the non-string command, path, and goal cases.
When LLMs return non-string parameter values (e.g. bool, int) in tool call
arguments, _summarize_tool_result() crashes with TypeError because it calls
len(), .count(), or slicing directly on args.get() return values.
This causes an infinite crash loop in the TUI — context compression
triggers on session resume, which crashes, which restarts, which triggers
compression again.
Add _str_arg() helper that coerces any value to str, and use it in all 5
vulnerable call sites:
- terminal: len(cmd)
- write_file: content.count()
- delegate_task: len(goal)
- execute_code: len(code)
- vision_analyze: question[:50]
* test: deflake async-delegation interrupt test under CI load
test_interrupt_all_signals_running_children failed twice on a loaded CI
worker: the blocker's ev.wait(timeout=5) expired before interrupt_all()
ran, the record finalized on its own, and interrupt_all() found nothing
running (n == 0, interrupted count 0 — the exact CI assertion failure,
reproduced locally by inserting a 5.6s dispatch->interrupt gap).
Raise the internal safety timeouts from 5s to 60s across the file's
gated runners — they exist only as runaway guards (every test releases
its gate explicitly); the pytest-level timeout is the real backstop.
Same flake class as the removed test_crashed_runner_produces_error_
completion (#64431).
* test: fix cross-test completion-event leak (second flake mechanism)
The first CI failure was the 5s guard-timeout race; the rerun exposed a
SECOND mechanism with a different signature: 'completed' == 'interrupted'.
Prior tests (e.g. test_dispatch_rejected_at_capacity) release their gate
and return immediately, but their workers finalize asynchronously — on a
loaded runner the teardown drain races the in-flight _finalize, and the
straggler 'completed' events leak into the NEXT test's queue, where
_drain_one() picks one up instead of the interrupt event. Reproduced
locally: gate release + immediate drain + 0.15s finalize delay leaked 2
events.
Fixes:
- teardown waits (bounded 2s) for active workers to finalize before
draining, so events land in the owning test
- the interrupt test matches its OWN delegation_id via _drain_for()
instead of taking whatever event arrives first
The 0.14 upgrade (#63970) dropped the @assistant-ui/store override, so the
whole cluster de-hoisted from root node_modules into apps/desktop/node_modules
under a single shared tap@0.9.3. The test only looked at the root hoist path
(node_modules/@assistant-ui/tap), which no longer exists, and failed on main.
Resolve tap wherever npm places it (root or workspace-nested), and assert a
single shared version across all install sites — strengthening the invariant to
also catch a split tap install, not just split declared ranges.
The assistant-ui upgrade lockfile omitted standalone @esbuild/* packages.
CI runs npm ci --ignore-scripts, so esbuild's postinstall never runs and
desktop/ui-tui check (build:ink, bundle-electron-main) fail without them.
Graft the 26 platform packages from main's lockfile.
@assistant-ui/react 0.14 makes respondToApproval required on
ToolCallMessagePartProps; the settledClarifyProps helper still lacked it
after the upgrade cherry-picks, so tsc failed on clarify-tool.test.tsx.
a reader of this subclass can't recover from hermes code alone that the metadata.isOptimistic flag drives core's off-branch eviction and export() omission, so a future core change to it would silently break placeholder cleanup. flagged in the upgrade review.
lock the four behaviors the built-in normalizeMathDelimiters/escapeCurrencyDollars introduce over the deleted custom helpers: $$<digit>$$ display math stays intact, double-backslash brackets and [/math]/[/inline] tag pairs rewrite to dollar delimiters, and currency dollars in prose are escaped. the existing preprocessMarkdown suite had no math cases.
the parseIncompleteMarkdown comment implied the reveal frontier is repaired; repair runs on the full accumulated text, so reword it to say that. drop the now-dead "multiple surfaces render the same content" clause from the block-cache comment (the smooth and defer wrappers that caused it were removed), and trim the math-preprocess comment to the load-bearing prose-only constraint.
@assistant-ui/store renamed its index-out-of-bounds throw from tapClientLookup/tapClientResource to useClientLookup in the 0.14 upgrade, so the boundary's /tapClient.../ filter stopped matching and re-threw the transient session-switch and reconnect race to root, blanking the app. broaden the regex to accept the new prefix (keeping the old one for older store versions) and point the test at the real message so it exercises the live path instead of the dead string.
delete the custom rewriteLatexBracketDelimiters and escapeCurrencyDollars
implementations from markdown-preprocess.ts (~40 lines). the built-in
exports from @assistant-ui/react-streamdown 0.3.4 are strict
improvements:
- normalizeMathDelimiters combines rewriteLatexBracketDelimiters (now
handles double backslashes and trims body whitespace) with
rewriteCustomMathTags (handles [/math]...[/math] and
[/inline]...[/inline] tags that some models emit — new capability
HA didn't have before)
- escapeCurrencyDollars excludes $ as a preceding character, so
display math $$5 is no longer incorrectly escaped (bugfix)
the call site in preprocessMarkdown changes from
rewriteLatexBracketDelimiters(escapeCurrencyDollars(part)) to
normalizeMathDelimiters(escapeCurrencyDollars(part)).
verified: tsc 0 errors, eslint clean, all 16 preprocessMarkdown tests
pass (including currency dollar escaping), vitest 0 new failures,
manual verification of currency amounts, LaTeX bracket delimiters,
display math, and dollar signs inside code blocks.
delete lib/remend-tail.ts (108 lines) and lib/remend-tail.test.ts (105
lines). the tailBoundedRemend export from @assistant-ui/react-streamdown
0.3.4 is algorithmically identical — same findRemendWindowStart boundary
scan, same fence/math tracking, same slice-and-repair strategy. the only
differences are improvements: the built-in handles \r (CR) in line
endings for Windows compatibility, and accepts an optional RemendOptions
parameter passed through to remend.
the import in markdown-text.tsx moves from @/lib/remend-tail to
@assistant-ui/react-streamdown. the call site
(preprocessWithTailRepair) is unchanged.
verified: tsc 0 errors, eslint clean, vitest 0 new failures (15
pre-existing, 786 passing — 6 fewer than before because the deleted
remend-tail.test.ts had 6 cases), manual verification of incomplete
markdown repair during streaming.
delete SmoothStreamingText, DeferStreamingText, and useSmoothReveal
(~174 lines) from markdown-text.tsx. the built-in defer and smooth
props on StreamdownTextPrimitive now handle the same work:
- defer: routes streaming text through useDeferredValue so markdown
re-parsing runs at lower priority (typing/scrolling stay responsive)
- smooth: typewriter-style reveal via useSmooth with SmoothOptions
{ drainMs: 500, maxCharsPerFrame: 30, minCommitMs: 33 }, matching
the old useSmoothReveal constants exactly
MarkdownTextContent (reasoning text) gets both defer and smooth.
MarkdownText (assistant text) gets defer only, matching the previous
behavior where text messages had no typewriter effect.
the internal pipeline order changes from smooth → defer → preprocess
to preprocess → smooth → defer (the built-in primitive runs preprocess
first). this is functionally equivalent: the tail-bounded remend repair
runs once on the full text instead of per revealed prefix, and the
smooth reveal operates on already-repaired markdown. end result is
identical.
verified: tsc 0 errors, eslint clean, vitest 0 new failures (15
pre-existing, 792 passing), manual verification of 6 streaming
scenarios (defer, smooth reveal, typing-while-streaming, code blocks,
math, long text performance).
bumps @assistant-ui/react from ^0.12.28 to ^0.14.23 and
@assistant-ui/react-streamdown from ^0.1.11 to ^0.3.4. this crosses
two minor bumps on each package and unlocks the built-in defer, smooth,
and tail-bounded remend primitives for PR 2.
breaking change from core 0.2.x: MessageRepository.appendOptimisticMessage
was removed (assistant-ui#4162). inline the three steps it did (generateId
+ fromThreadMessageLike + addOrUpdateMessage) in
incremental-external-store-runtime.ts, and set metadata.isOptimistic so
the new off-branch eviction logic cleans up the placeholder correctly.
fromThreadMessageLike and generateId graduated to the public API in
0.14.22 (assistant-ui#4414), so they now import from @assistant-ui/react
instead of @assistant-ui/core/internal. ExportedMessageRepository in the
test file moves to the public import for the same reason. the remaining
internal imports (AssistantRuntimeImpl, BaseAssistantRuntimeCore,
ExternalStoreThreadListRuntimeCore, ExternalStoreThreadRuntimeCore,
hasUpcomingMessage) are runtime construction internals with no public
equivalent and stay on @assistant-ui/core/internal.
the @assistant-ui/store npm override is removed: all transitive ranges
now resolve to 0.2.18 without it.
verified: tsc --noEmit passes, vitest shows zero new failures (15
pre-existing, 792 passing, identical to baseline before the upgrade).
The chat_completions path raises EmptyStreamError when a stream yields
no chunks and no finish_reason (#64420). _call_anthropic() had no
equivalent, and the eventless failure surfaces differently by client:
- Real Anthropic SDK: no message_start means no final-message snapshot,
so get_final_message() raises a bare AssertionError — not in the
retry loop's transient set, so it burned no retries and surfaced raw.
- OpenAI-compat shims: may fabricate a contentless Message with no
stop_reason (or return None), flowing out as a 'successful' empty turn.
Track whether any stream event arrived and normalize both shapes to
EmptyStreamError, giving the anthropic path the same transient retry
budget in the shared _call() retry loop. A real completed response
always carries a stop_reason, so the guard cannot fire on legitimate
turns (including eventless mocks with stop_reason='end_turn').
Follow-up to #64420.
An exhausted EmptyStreamError previously reported 'Connection to provider
failed' — misleading, since the connection succeeded (stream opened) and
the provider simply sent nothing. Add a dedicated third branch so users
debugging a misconfigured endpoint aren't sent chasing network issues.
Follow-up to #64420.
Three-layer companion to the salvaged CLI drain-ownership fix (#64240):
1. restore_undelivered_completions stamps restored=True (in-memory only)
on every durable completion re-enqueued at process start.
2. drain_notifications' legacy unfiltered branch re-queues restored
events instead of consuming them — a fresh process can no longer
adopt a dead session's delegation results (#64484). Same-process
keyless events keep the legacy behavior.
3. delegate_tool's async dispatch now falls back to the parent agent's
durable session_id when the approval-context key resolves empty (the
CLI case), so the CLI's new positive-ownership drain can actually
claim its own completions instead of failing closed on ''.
Review finding on the salvaged collector: _wait_for_process is the shared
drain for EVERY env.execute() consumer, not just the terminal tool. Applying
tool_output.max_bytes there silently truncated file-operation cat reads
(read_file_raw feeds the patch engine — read-modify-write on any file >50KB
would corrupt it), paginated read_file, code-execution RPC reads, and log
reads.
bounded_capture is now an explicit opt-in on execute()/_wait_for_process,
set only by the foreground terminal tool. Default preserves the historical
full-fidelity capture via an effectively-unbounded collector (single code
path). Modal transports accept the kwarg for signature parity.
New regression test: default execute() returns a 200KB payload complete and
untruncated. E2E: 20MB internal read intact; ShellFileOperations
read_file_raw round-trips byte-exact; terminal path still bounded at 50KB.
Follow-up to @michaelHMK's cherry-picked fix for #50892:
- agent/tool_dispatch_helpers.py: guard the segment planner's except-path
debug log against non-string arguments (the planner replaced the old
_should_parallelize_tool_batch body after #64460 and inherited the same
latent arguments[:200] slice).
- tests: the PR's executor-coercion hunks were dropped as redundant —
both executors now route through _parse_tool_arguments, which already
rejects null/non-object args with a structured error result instead of
coercing to {} (the 'we do not repair bad model outputs' contract).
Reworked the salvaged tests to pin the current behavior: None args are
rejected without dispatch, valid siblings still run, the planner treats
them as a barrier without raising, and the mainline run_conversation
path (which normalizes None to '{}' before dispatch) stays crash-free
under verbose logging.
Review nit from verification: the comment claimed newer FastMCP accepts a
JSON schema kwarg — the installed SDK's add_tool has no such parameter; the
synthesized __signature__ is what drives schema generation on both paths.
Fixes#64025
The hermes-tools MCP server was fetching each tool's JSON schema
into params_schema but never passing it to FastMCP's add_tool(),
so all published tools had an empty **kwargs signature. MCP clients
couldn't see parameters and arguments were dropped at dispatch.
This fix:
- Adds _signature_from_schema() to convert JSON schemas to Python
function signatures with type annotations
- Attaches the generated signature/annotations to each handler closure
so FastMCP introspects the real parameter structure
- Filters out None values before dispatch to avoid forwarding unset
optional parameters
Impact: web_search, browser automation, vision, and other Hermes tools
are now properly callable from the codex_app_server runtime.
Salvaged from #32340 by @LeonSGP43, adapted to the workspace-scoped
status tracking that landed in #63709:
- /stop with no running agent now best-effort clears the platform
status indicator, so a phantom 'is thinking...' left by a gateway
restart or a turn that died without a final send can always be
dismissed (#32295).
- SlackAdapter.stop_typing clears an untracked thread when the caller
names it explicitly in metadata — clearing an unset status is a
harmless no-op on Slack's side. The fallback is skipped when multiple
Slack Connect workspaces track the same channel+thread and no team_id
is given, preserving #63709's cross-workspace safety guarantee.
The upstream migration to inference.nousresearch.com broke routing for
inference-api.nousresearch.com. Restore the correct base_url and drop
the stale alias match in _is_nous_inference_route().
Fixes#60715
The read-only premise only holds on Python 3.13+ where the full PTB request
MRO is slotted. CI runs 3.11/3.12 where BaseRequest instances still carry a
__dict__, so the unconditional pytest.raises failed. The behavioral half of
the test (subclass re-tag instruments and records) runs everywhere.
Adds a __slots__ request double that has no instance __dict__, mirroring
PTB's HTTPXRequest shape on Python 3.13. The test asserts the old
instance monkey-patch is rejected as read-only and that
_instrument_polling_request instead re-tags the class and still records
getUpdates progress. Fails on the pre-fix adapter with the exact
"'do_request' is read-only" AttributeError from the report.
PTB's HTTPXRequest/BaseRequest use __slots__. On Python 3.13 their
instances no longer carry a __dict__, so the getUpdates progress
instrumentation's `request.do_request = wrapper` monkey-patch raises
`AttributeError: 'HTTPXRequest' object attribute 'do_request' is
read-only`, failing every Telegram connect (#64482). It only worked on
Python 3.12, where the instance still had a __dict__ — the Docker image
ships Python 3.13, so the release broke Telegram outright.
Re-tag the request to a thin `__slots__ = ()` subclass that overrides
do_request instead of mutating the instance. This preserves the exact
progress-observation semantics, works on both 3.12 and 3.13, and covers
the real request and the test doubles alike.
When a user starts a chat without ever selecting a model (GUI Chat App
onboarding, provider-set-but-model-missing config, empty model.default),
every silent fallback path resolved to the first curated catalog entry —
anthropic/claude-fable-5, the priciest flagship. Users were silently
billed for the most expensive model without opting in.
- hermes_cli/models.py: add PREFERRED_SILENT_DEFAULT_MODEL (z-ai/glm-5.2)
+ pick_silent_default_model() helper; point the nous silent-default
override at it and add an openrouter override (previously resolved to
"" and let downstream paths land on the flagship).
- hermes_cli/web_server.py: /api/model/recommended-default (the endpoint
the Desktop onboarding confirm card reads) now picks GLM-5.2 when the
provider's list carries it instead of blindly taking entry [0].
- tui_gateway/server.py: _resolve_model()'s last-resort literal was
anthropic/claude-sonnet-4; now PREFERRED_SILENT_DEFAULT_MODEL.
- tests: update empty-model fallback tests for the new contract.
The just-merged auxiliary.<task>.reasoning_effort shorthand applied
ensemble-wide to MoA (one value for every advisor) — wrong granularity.
Per-slot preset config supersedes it:
moa:
presets:
deep_review:
reference_models:
- {provider: ..., model: ..., reasoning_effort: low}
- {provider: ..., model: ..., reasoning_effort: xhigh}
aggregator:
{provider: ..., model: ..., reasoning_effort: high}
- Remove reasoning_effort from the moa_reference/moa_aggregator
DEFAULT_CONFIG blocks; _get_task_extra_body now warns-and-ignores the
key on MoA tasks, pointing at the preset config
- Guard tests: MoA aux blocks must not regrow the key; task-level value
is rejected with the pointer warning
- Docs: configuration.md notes the MoA exception and links the MoA page
Add patterns 30-34: forced metaphors/figurative overwriting, dramatic fragmentation and punchy kickers, rhetorical questions answered immediately, sentence-opener tics, and reassurance kickers. These cover structural/rhythm tells the existing 29 patterns miss (short subjectless dramatic fragments, cutesy aphorisms, rhetorical Q-and-A).
Add a 'marketing and blog cliches' list to pattern 7 for the business/LinkedIn register (game-changer, circle back, deep dive, moving forward, ...) that the Wikipedia-derived vocabulary list does not catch.
Edit the skill's own instructional prose to follow its own guidance: remove em dashes and negative parallelism from the narration so the skill models the writing it asks for. Before/After examples are left intact so the demonstrations still show the tells.
Update the pattern count (29 -> 34) and the attribution note to mark 30-34 and the pattern-7 additions as Hermes additions beyond the blader/humanizer source.
Terminal and logs now follow the IntelliJ/VS-Code tool-window model: their
toggle (⌃`, ⌘K) COLLAPSES the zone to a rail with the tab still showing instead
of hiding it outright, so "toggle" and "the tab bar" stop fighting. Restore
routes through the pane's store opener (rail click / chevron) so the shortcut and
titlebar toggle stay truthful; the tab's ✕ dismisses the panel (comes back via
its toggle), while a session tile's ✕ still closes the session. New store
primitives (setPaneCollapsed / restoreTreePane / collapseTreePane + a
collapse-pane registry) via a bindPaneCollapse in the controller.
Phase 3 of relay-channel-context (gateway/agent side, single PR). The
connector (gateway-gateway #122/#123/#124) now attaches read-only
surrounding channel/group context to an addressed relay turn; this wires
the gateway to consume it.
- descriptor.py: additive optional supports_context (default False) on
CapabilityDescriptor. from_json already filters unknown keys, so this is
back-compat both directions within contract_version 1.
- ws_transport.py: _event_from_wire maps the connector's read-only
context[] array into the EXISTING MessageEvent.channel_context field via
a new _render_relay_context() helper — reusing the same read-only
injection path history-backfill uses (run.py prepends channel_context
ahead of the trigger message). Never raises; absent/empty/malformed ->
channel_context unset (byte-identical to today).
- docs/relay-connector-contract.md: document supports_context in the §2
descriptor table (fixes the contract-doc conformance test) + the
context/context_error inbound fields in §3.
- tests: descriptor default/round-trip/forward-compat; _render_relay_context
rendering + malformed-safe; _event_from_wire context->channel_context
mapping + the read-only invariant (trigger text untouched).
RELAY-ONLY: only gateway/relay/* + the shared MessageEvent consumption via
its existing channel_context field. No native adapter touched.
The inbound cron-fire verifier constructed a fresh PyJWKClient on every
fire, discarding the client's key cache and forcing a synchronous JWKS
HTTP GET to the portal on each fire. Under a burst of concurrent fires
(a hosted instance with several cron jobs firing in the same window) this
fanned out into N simultaneous JWKS fetches that the portal rate-limited
(HTTP 403 -> verification fails -> agent 401), or that blocked the event
loop long enough that the fire webhook could not return its 202 before
the relay's 30s timeout (observed in prod as relay 504s concentrated on
high-job-count instances).
Cache one PyJWKClient per JWKS URL at module scope (double-checked lock)
so the signing keys are reused across fires; NAS keys rotate rarely, so
the steady state is zero JWKS fetches per fire.
Regression test proves 5 fires -> 1 client construction (was 5).
The incremental external-store runtime reconciles message repositories in place
(addOrUpdateMessage + prune-non-incoming). On a session switch the incoming
transcript shares no ids with the current one, and grafting the new chain onto
the old tree before pruning can strand a stale head/branch — the thread keeps
showing the previous session. When nothing carries over there's nothing to
preserve, so clear the tree first (leaves→root) then rebuild clean. Belt-and-
suspenders alongside the $messages-carryover fix.
Every auxiliary task block (vision, web_extract, compression,
title_generation, curator, background_review, moa_reference, ...) now
accepts a reasoning_effort shorthand:
auxiliary:
compression:
reasoning_effort: low
vision:
reasoning_effort: none
_get_task_extra_body() folds it into extra_body.reasoning, which every
auxiliary wire already translates: chat.completions passes it through,
the Codex Responses adapter maps it to top-level reasoning/include, and
the Anthropic auxiliary adapter now forwards it into
build_anthropic_kwargs(reasoning_config=...) (previously hardcoded None).
An explicit extra_body.reasoning on the same task wins over the
shorthand. Invalid levels are ignored with a warning. Empty string
(the shipped default) is a no-op — zero behavior change.
Config: reasoning_effort added to all 16 auxiliary task blocks in
DEFAULT_CONFIG (no version bump — deep-merge handles new keys).
Follow-ups on top of #64061's salvage:
- ResourceLink markers now point at mcp__<server>__read_resource (the
actual registered tool name via mcp_prefixed_tool_name) instead of a
nonexistent <server>_read_resource the agent could hallucinate-call.
- The isError path now surfaces EmbeddedResource .resource.text blocks
instead of dropping them, so error payloads carried in resources no
longer collapse to a bare 'MCP tool returned an error'. (Same-class
fix flagged in #64061 and independently addressed in #63576 by
@alauer.)
- 3 new error-path tests + updated ResourceLink wire-name assertion.
MCP tool results with non-image binary resources (PDFs, archives, office
docs) were silently dropped: the success path only handled TextContent and
ImageContent, so a PDF-returning MCP tool appeared to return metadata only.
- EmbeddedResource blob contents are decoded (50MB cap), materialized into
the Hermes document cache via cache_document_from_bytes (sanitized
filename, traversal-safe), and surfaced as a local-path marker the agent
can read with file/terminal tools.
- EmbeddedResource text contents are inlined directly.
- ResourceLink blocks preserve the URI and point the agent at the server's
read_resource tool; no arbitrary network fetch outside the MCP session.
- AudioContent blocks are cached via cache_audio_from_bytes as MEDIA: tags.
- read_resource blob contents are materialized the same way instead of
returning '[binary data, N bytes]'.
- Unsupported blocks are logged instead of silently discarded.
- Existing ImageContent MEDIA: behavior unchanged.
Reported by an enterprise customer; reproduced against an HTTP MCP server
returning application/pdf resources.
Slack's June 30 Agent messaging experience changelog lists Bolt Python
1.29.0 / Python SDK 3.43.0 as the Agent View minimums. Bump the
messaging/slack extras and the platform.slack lazy-install pins to
match, and regenerate uv.lock. All adapter API surfaces verified
present against the new versions in a clean venv.
The docs state feedback_buttons requires rich_blocks: true, but
_maybe_blocks rendered full Block Kit whenever feedback_buttons alone
was enabled — implicitly turning on rich-block rendering the user never
opted into. Align the code with the documented contract and add a
regression test.
The faint statue backdrop behind the transcript was only switchable via
the DEV-only leva panel. Add a persisted Appearance toggle (default on)
so users can hide it; the Backdrop simply skips rendering when off.
resumeSession hand-rolls $messages (it paints before a runtime id is bound), and
only cleared the old transcript on the cold path at entry. But a warm-cache hit
can bail down to the full resume — an empty-transcript drop, or the cache being
purged during the profile-swap await — without ever clearing, so the previous
session's array leaked into the next one. Symptom: switching sessions kept
showing the same messages (deterministic once tiling pre-warms the cache on
boot). Clear $messages at the single point every cold/bail path converges, so
carryover is structurally impossible; the warm fast-path still repaints in place.
The terminal environment is shared process-globally (collapsed to the
default key), so env.cwd tracks the LAST session that ran a command.
_resolve_command_cwd() trusted env.cwd unconditionally — no ownership
check — so when session A left env.cwd pointing at A's checkout,
session B's first terminal command inherited A's stale cwd and ran in
the wrong workspace.
The file tools already solved this exact shared-env problem with
_live_cwd_if_owned() checking env.cwd_owner. The terminal tool never
got the same guard.
Fix: capture env.cwd_owner BEFORE the current session claims it, and
pass it as prev_owner to _resolve_command_cwd. When the previous owner
was a different session, env.cwd is stale — fall through to default_cwd
(the config/override cwd for this session) instead. Once the session
has claimed the env, subsequent calls in the same session still trust
env.cwd so in-session state survives.
A model response containing several parallel-safe reads plus one unsafe
tool used to lose ALL concurrency: _should_parallelize_tool_batch was
all-or-nothing, so a single barrier call (terminal, clarify, unknown
tool, malformed args) forced the entire batch onto the sequential path.
_plan_tool_batch_segments now splits the batch into ordered segments:
maximal contiguous runs of parallel-safe calls execute on the existing
concurrent path, barrier calls on the sequential path, strictly in the
model's emission order. Invariants preserved:
- one tool result per call, appended in emission order (segments are
contiguous, so no result reordering across a barrier)
- side-effect boundaries: no call starts before an earlier barrier ends
- overlapping file targets split into separate ordered parallel runs
- turn-end budget enforcement + /steer injection run exactly once per
batch (segment executors run with finalize=False; the segmented
dispatcher owns the whole-turn finalize)
- interrupt during segment k drains segments k+1..n with cancelled
results, keeping one result per tool_call_id
Homogeneous batches keep their original single-path dispatch (zero
behavior delta); _should_parallelize_tool_batch remains as a thin view
over the planner for existing callers and tests.
Two test mocks stubbed the old zero-arg signature; the chokepoint refactor
added an optional model param that call sites now pass. Swept the full test
tree for other stale stubs of the changed functions — the rest use
MagicMock/patch(return_value=...), which tolerate the new arg.
Collapse the six per-surface copies of override-then-global resolution
(CLI startup, gateway, TUI, cron, /model switch, fallback activation)
onto one shared resolve_reasoning_config() in hermes_constants.
Also fixes the gateway resolving reasoning against config model.default
instead of the session's effective model: after a session-only /model
switch, the switched model's override now applies (gateway message paths
pass the resolved session model through _resolve_session_reasoning_config;
/reasoning status reads the session model override).
Cleanup: drop docs/PER_MODEL_REASONING.md (duplicates the website docs
page), drop the change-detector _config_version test (no bump needed —
deep-merge handles new keys), remove a stale plan-reference comment.
Adds chokepoint contract tests (13) and gateway session-effective-model
regression tests (2).
Add agent.reasoning_overrides dict to config.yaml. Users can now set
a reasoning_effort per model, overriding the global agent.reasoning_effort.
Example:
agent:
reasoning_effort: "medium" # global default
reasoning_overrides:
"openrouter/anthropic/claude-opus-4.5": "xhigh"
"openai/gpt-5": "low"
"claude-sonnet-4.6": "high" # bare model name also works
The helper is spelling-tolerant: override keys match regardless of
provider prefix or dots-vs-dashes normalization, so users can write
keys in any sensible form and they'll match.
Resolution priority:
1. Session-scoped /reasoning --session override (gateway only; unchanged)
2. Per-model override from agent.reasoning_overrides (spelling-tolerant)
3. Global agent.reasoning_effort (existing)
4. Provider default (unchanged)
Wired into:
- CLI startup (cli.py)
- Messaging gateway agent construction (gateway/run.py)
- Desktop/TUI _load_reasoning_config (tui_gateway/server.py)
- Cron job scheduler (cron/scheduler.py)
- /model mid-session switch (agent/agent_runtime_helpers.py)
+ _primary_runtime now tracks reasoning_config for correct fallback recovery
- Fallback activation (agent/chat_completion_helpers.py::try_activate_fallback)
+ Re-resolves reasoning_config for the fallback model (best-effort)
Closes#21256 (per-model reasoning_effort defaults).
Note: no hermes config set agent.reasoning_overrides.<model> support;
users edit the YAML directly. _set_nested splits on "." and would
corrupt model keys containing version dots.
Review findings from the 4-angle pass:
- Unknown-but-enabled effort levels now collapse to Solar's strongest
(high) instead of silently downgrading to the medium default — guards
against the next #62650-style vocabulary addition. Explicit-empty
effort keeps the medium default.
- fallback_models test now asserts the behavior contract (non-empty, no
denied families) instead of freezing the exact model tuple
(change-detector, AGENTS.md reject reason).
- Drop unused pytest import in test_upstage_provider.py.
Main added max/ultra effort levels (#62650) after this PR branched;
without the mapping 'ultra' silently fell through to the medium default.
Matches the xhigh/max collapse-to-strongest convention used by other
profiles.
PROVIDER_REGISTRY, its alias map, and CANONICAL_PROVIDERS all auto-extend
from registered ProviderProfiles since the provider-modules refactor
(20a4f79ed). Verified with real imports: registry entry, 'solar' alias
resolution via resolve_provider(), and the picker entry are identical
with the manual entries removed. The hermes_cli/providers.py overlay
stays (models.dev has a stale /v1/solar base URL and no UPSTAGE_BASE_URL
var), and the manual OPTIONAL_ENV_VARS entries stay (non-advanced key +
curated prompt text, matching the fireworks convention).
Pin the Upstage default to the concrete solar-pro3 instead of the
solar-pro rolling alias:
- plugin fallback_models is now ("solar-pro3",); entry [0] is the setup default
- drop the "solar-pro" context-window fallback entry (solar-pro3 covers it)
- update the reasoning default-on docstring and profile tests accordingly
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Remove the `solar-open2-preview` context-window entry; `solar-open2`
covers the Open 2 family at the same 256K window.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
`UPSTAGE_API_KEY` / `UPSTAGE_BASE_URL` were wired through the provider
resolver, auth registry, and the EnvPage grouping, but never added to
`OPTIONAL_ENV_VARS` in hermes_cli/config.py. The dashboard/desktop
Providers page builds its list from that catalog (`/api/env` iterates
`OPTIONAL_ENV_VARS`), so with no entry the keys were never emitted and
"Upstage Solar" never rendered — the EnvPage prefix group stayed empty.
Add both keys under `category: "provider"` (matching gmi/minimax) so they
show up in `hermes dashboard` / `hermes desktop` under "Upstage Solar".
Adds a regression test asserting the catalog contains them, mirroring the
existing GMI coverage.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Invert the reasoning-support check from an allow-list (solar-pro,
solar-open) to a deny-list of the known non-reasoning families
(solar-mini, syn-pro). Newly released Solar models now get
reasoning_effort by default instead of having it silently dropped.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Adds Upstage Solar as a bundled model-provider plugin. Solar exposes an
OpenAI-compatible chat-completions endpoint at https://api.upstage.ai/v1, so
the generic chat_completions transport handles request/response/streaming/tool
calls — the profile is the core integration.
Provider registration (Upstage isn't in models.dev, so each registry that does
not auto-wire from the plugin layer needs an explicit entry — same pattern as
nvidia/gmi):
- plugins/model-providers/upstage/: UpstageProfile + plugin.yaml. Picker default
and offline catalog list only the agentic Solar Pro models, led by `solar-pro`
(rolling alias for the latest Pro). default_aux_model empty so aux tasks use
the main model. `solar` alias. UPSTAGE_BASE_URL overrides the host.
- hermes_cli/providers.py: HERMES_OVERLAYS + label + `solar` alias, so
resolve_provider_full('upstage') resolves (without this, an explicit
`provider: upstage` in config was dropped and fell through to auto-detect).
- hermes_cli/auth.py: PROVIDER_REGISTRY entry + `solar` alias, so `hermes
doctor` / resolve_provider recognise upstage (the static-registry path the
lazy profile-extension doesn't reliably cover at validation time).
- hermes_cli/models.py: CANONICAL_PROVIDERS entry places Upstage Solar in the
curated picker order (above the auto-appended `custom`).
- agent/model_metadata.py: context-window fallbacks (/v1/models omits
context_length); `solar-pro` carries the 128K Pro context as the catch-all.
Reasoning: UpstageProfile.build_api_kwargs_extras wires Solar's top-level
`reasoning_effort` (low|medium|high; xhigh/max→high). Reasoning-capable families
are solar-pro* and solar-open*; solar-mini/syn-pro never receive it. Defaults ON
at medium when unset (matches the /reasoning "medium (default)" label);
`/reasoning none` disables; explicit/saved settings are honored. No
reasoning_content echo handling needed (unlike DeepSeek/Kimi).
Web dashboard:
- web/src/pages/EnvPage.tsx: add an "Upstage Solar" provider group so
UPSTAGE_API_KEY / UPSTAGE_BASE_URL appear under LLM Providers (not "Other").
Docs/tests:
- .env.example: documents UPSTAGE_API_KEY / UPSTAGE_BASE_URL.
- tests: profile wiring, reasoning_effort mapping (pro/open/mini, efforts,
disabled, default-on), provider-resolver regression (resolve_provider_full /
get_provider / solar alias / overlay), `solar-pro` default.
Testing: pytest tests/providers tests/plugins/model_providers
tests/hermes_cli/test_upstage_provider.py tests/run_agent/test_provider_parity.py
tests/hermes_cli/test_api_key_providers.py; ruff clean. Verified end-to-end:
`hermes doctor` shows "Upstage Solar", and live chat works via both
`--provider upstage` and `--provider solar`. Reasoning wire format per
https://console.upstage.ai/api/docs/for-agents/raw. Platforms tested: macOS.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Follow-up to the cherry-picked empty-user-turn drop: the placeholder
introduced in 8582f35d9 fired for whitespace-only STRING turns too
(content=' ' flattens to non-stripping text but isn't in the
(None, '', []) exclusion set), fabricating an attachment note for a turn
that carried nothing. Gate the placeholder on isinstance(content, list)
so only genuinely structured (e.g. image-only) turns get it; empty and
whitespace-only string turns now fall through to the drop path.
Edge cases verified: trailing empty user turn still ends the view on the
synthetic advisory marker; an all-empty transcript degenerates to [].
MoA's _reference_messages() unconditionally appended every user-role
message to the advisory view sent to reference models, even when the
message content was an empty string or a non-string/multimodal payload
that the text-extraction step flattens to "".
Strict providers (Kimi/Moonshot, and others that enforce non-empty user
content) reject such a message with:
400 Invalid request: the message at position N with role 'user'
must not be empty
Lenient providers (DeepSeek) accept it, so an identical rendered view
passes on one reference and 400s on another within the same fan-out —
the user sees "kimi doesn't support MoA" when the real cause is an empty
user turn leaking into the advisory transcript.
Skip empty user turns, mirroring the existing behavior for empty
assistant turns (which are already dropped when they carry no parts).
The end-on-user invariant is preserved: the synthetic advisory-request
user turn is still appended when the view would otherwise end on an
assistant turn.
Adds a regression test asserting the advisory view contains no empty
user turn and still ends on a user turn.
After #53825's fix removed the auto-rich table bypass from
_rich_delivery_enabled(), _content_is_pipe_table_primary() had zero
callers. Remove it and simplify _rich_delivery_enabled() to the bare
rich_messages opt-in check (content param no longer used).
The platform hint in PLATFORM_HINTS['telegram'] always encouraged rich
Markdown constructs (tables, task lists, math, collapsible details) even
when rich_messages: false (the default). This caused the agent to produce
formatting that MarkdownV2 cannot render, especially broken on Telegram Web.
Split the hint into a base hint (MarkdownV2-compatible) and a
TELEGRAM_RICH_MESSAGES_HINT extension. The extension is conditionally
appended in system_prompt.py only when
platforms.telegram.extra.rich_messages is true.
Fixes#57122
Remove the pipe-table bypass from _rich_delivery_enabled() so that
rich_messages: false is fully honoured. Previously, pipe tables were
auto-routed to sendRichMessage regardless of the config flag, breaking
delivery on clients without Bot API 10.1 support (AyuGram, Telegram
Web, some desktop clients).
Fixes#53824
Flaked 3 times today across 3 unrelated PRs (#64321, #64319, #64409),
on two different CI shards (slice 1 and slice 8), while passing
deterministically on local runs of the same SHAs. The test polls
process_registry.completion_queue for 5s waiting for a daemon-thread
completion event; since the durable completion delivery work
(67f4e1b4a, d0e9a42ce) the crashed-runner path also writes through the
sqlite-backed persistence layer, and on slow CI runners the in-memory
enqueue can lose the 5s race.
Coverage note: the durable-delivery suite in this file covers the
completed-runner and submit-failure paths through persistence, but not
a runner that raises mid-flight — that specific path loses its direct
test with this removal. A deterministic (non-racing) replacement can
follow separately if wanted.
* fix(desktop): clear stale compaction status
Clear the compaction phase when a turn resumes with model or tool activity, and key response timers by session and turn so switching chats preserves elapsed time.\n\nSupersedes #48115 by porting its resumed-content approach to the current split stream hook and covering tool-first resumptions.\n\nCo-authored-by: liuhao1024 <sunsky.lau@gmail.com>
* fix(desktop): resume after thinking activity
* fix(desktop): clear turn timer on stop
CI test slices don't install python-telegram-bot (optional dep), causing
a ModuleNotFoundError on collection. Add pytest.importorskip('telegram')
before the PTB imports.
Review follow-up to the reasoning_config cache-parity fix:
- Only inherit the parent's reasoning_config when the fork runs on the
parent's model (not routed). On the routed aux path
(auxiliary.background_review.{provider,model}) the cache is cold
regardless, so parity buys nothing, and the parent's effort vocabulary
can be invalid for the routed model/provider: OpenRouter
extra_body.reasoning.effort is forwarded unclamped
(chat_completions.py) and codex_responses only maps max/ultra for
gpt-5.6 — an exotic parent effort routed to a strict provider could
400 the review. Mirrors the existing 'not _routed' gate on
_cached_system_prompt / session_start three lines below.
- Add a routed-path regression test asserting reasoning_config is
omitted from the fork kwargs when _resolve_review_runtime returns
routed=True.
- Extract the four copy-pasted recorder stubs in
test_background_review_cache_parity.py into a single
_make_recorder_class() factory so a new fork attribute needs one stub
edit, not four.
PR #17276 painstakingly pinned `_cached_system_prompt`, `session_start`,
`session_id`, and the toolset config on the background-review fork so its
outbound request body would byte-match the parent's and hit Anthropic's
exact-prefix cache. The contributor measured a ~26% end-to-end cost
reduction on Sonnet 4.5.
That optimization is currently being silently undone by a missing
`reasoning_config` kwarg. The fork's `AIAgent(...)` call omits it, so the
fork's `reasoning_config` defaults to `None`. `anthropic_adapter.build_anthropic_kwargs`
(line ~2165) then short-circuits the `thinking` / `output_config` block,
and the fork's request body lands in a DIFFERENT Anthropic cache namespace
from the parent's.
Result on the wire: 0 `cache_read_input_tokens`, full `cache_creation_input_tokens`
of the entire parent prefix — every single background review.
7 days of midagent.db traffic from one host running stock Hermes against
Anthropic Sonnet:
```
Background-review FIRST calls (the moment a review fork is born):
count = 68
cache_write tokens = 7,004,297
cache_read tokens = 1,016,335
Cost on Sonnet ($3.75/M write vs $0.30/M read):
Spent on these writes: $26.27
Cost if they had hit parent cache instead: $2.10
WASTED: $24.16 / week / user
```
That is from one user. Multiply by Hermes's installed base for the full
impact.
Tested against api.anthropic.com directly (see refs/api-tests/ in the
attached investigation repo if needed):
| pair | cache_r | cache_w |
|---------------------------------------------|---------|---------|
| parent fresh | 0 | 24,047 |
| parent same again | 24,047 | 0 |
| fork: appends 2 new tail msgs, thinking ON | 24,047 | 22 |
| fork: appends 2 new tail msgs, thinking OFF | 0 | 24,047 |
Same fork-shape request, only difference is `thinking`. With the fix,
the fork hits the parent's full prefix and only writes the delta
(the `Review the conversation above…` prompt block, ~3-5K tokens).
One line in `agent/background_review.py`: pass
`reasoning_config=getattr(agent, "reasoning_config", None)` to the
`AIAgent(...)` constructor of the review fork. A short comment block
above it explains why so the next person who reads this code doesn't
re-introduce the regression.
`tests/run_agent/test_background_review_cache_parity.py` already covers
the system-prompt / session-id / toolset-config parity contracts that
PR #17276 introduced. I added:
* a `reasoning_config` attribute to `_make_agent_stub` so the stub has
a non-None parent value the test can verify is propagated.
* `test_review_fork_inherits_parent_reasoning_config()` — asserts the
fork's `AIAgent(...)` kwargs carry the parent's `reasoning_config`.
Pre-fix this test fails with `None vs expected {'enabled': True, 'effort': 'medium'}`;
post-fix all 4 tests in the file pass.
```
$ python -m pytest tests/run_agent/test_background_review_cache_parity.py -v
test_review_fork_inherits_parent_cached_system_prompt PASSED
test_review_fork_pins_session_start_and_session_id PASSED
test_review_fork_inherits_parent_toolset_config PASSED
test_review_fork_inherits_parent_reasoning_config PASSED ← new
```
Also runs against the broader background-review test suite:
`test_background_review.py` (4), `test_background_review_summary.py` (8),
`test_background_review_toolset_restriction.py` (3) — 19/19 pass.
`agent/curator.py:1691` has the same omission for the umbrella-curation
fork, but curator's prompt is "curate all skills" — it shares no prefix
with any user conversation, so cache-parity is a non-issue there. Worth
auditing if the curator ever takes a parent conversation as input, but
not part of this PR.
The `agent/auxiliary_client.py:1006` `reasoning_config=None` hardcode is
intentional (title/summary one-shots on short prompts — per-call cost
of namespace flip is negligible) and is also out of scope.
prune_old_entries' active-process guard failed open: when
has_active_processes_fn raised, the except block logged at debug and
fell through to the age check, so sessions with live background
processes attached could still be pruned — violating the documented
invariant that such sessions are never dropped. Add a continue so an
exception in the safety check fails safe (the entry is kept).
Commit 6b408e131 fixed the session_key/session_id mismatch in this
same guard but left the exception path failing open.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The #63309 hang class — gateway stuck at 'Connecting to Telegram
(attempt 1/8)' with no retry, no timeout, for minutes — can only occur
when the event loop thread itself is blocked in a synchronous call:
_await_with_thread_deadline's timer fires off-loop, but its expiry
hand-off (call_soon_threadsafe) still needs the loop to run, and the
gateway's outer wait_for is a pure loop timer. When the loop is pinned,
every layer goes silent simultaneously and the process wedges with no
evidence of where.
Two changes:
1. Loop-blocked watchdog in _await_with_thread_deadline: a second
daemon timer fires one grace period (5s) after the deadline; if the
loop still hasn't processed the expiry, it logs a WARNING from the
timer thread and faulthandler-dumps all thread stacks to stderr —
converting the silent hang into a trace that names the exact
blocking frame. A threading.Event set by the expiry callback (and on
normal exit) keeps completed awaits from ever being misreported.
2. discover_fallback_ips: the system-resolver leg runs
socket.getaddrinfo in a worker thread with no timeout, and
asyncio.gather waited on it unboundedly — a wedged OS resolver
stalled discovery for minutes between the two startup log lines. Its
result only feeds a log message, so it no longer gates discovery:
DoH legs (already client-bounded) are gathered alone and the system
leg is awaited with a _DOH_TIMEOUT cap, best-effort.
Refs #63309
Tests: 3 watchdog regressions (blocked-loop dump fires; responsive-loop
timeout does not; completed await does not) + 2 hung-resolver
regressions (DoH results returned promptly; worst-case seed fallback
stays bounded).
Exercises the actual ShellFileOperations.write_file and patch_replace
code paths (not just the helper in isolation) to verify that
safe-root denials surface 'outside HERMES_WRITE_SAFE_ROOT' and
credential-path denials surface 'protected system/credential file'.
Adapted from PR #55615 by @liuhao1024.
Return actionable errors when HERMES_WRITE_SAFE_ROOT blocks a path instead of
labeling every denial as a protected credential file. Wire the helper through
write_file, patch, delete/move, and the Copilot ACP shim; sync docs examples.
Document that safe-root violations are hard-blocked (not approval-gated),
add a security guide section for write_file/patch guards, and link cron
and verifier docs so users trust the footer over agent summaries.
Regression tests for find_latest_gateway_session_for_peer and
SessionStore stale-routing self-heal when end_reason is ws_orphan_reap.
Pin manual approval mode in blocking E2E tests so smart aux-LLM
resolution does not flake CI.
Whitelist ws_orphan_reap alongside agent_close in
find_latest_gateway_session_for_peer so gateway stale-routing
self-heal can reopen wrongly-reaped messaging sessions instead of
minting empty replacements. Layer A prevention already landed in #60609.
Review round 2 from @ethernet8023 on #61580:
1. The manifest list was a hardcoded root/ui-tui/web trio — desktop and
any future workspace escaped the skip key even though step 1's root
install hoists deps for every workspace. The list is now expanded
from the root package.json 'workspaces' globs (npm's own source of
truth): on the real repo that yields all 8 manifests incl.
apps/desktop, apps/bootstrap-installer, apps/shared, and the nested
ui-tui/packages/hermes-ink. Unreadable package.json falls back to
root manifests only (never skips more than main would install).
2. --prefer-offline dropped entirely (this branch no longer carries
#39399): local 3-run benchmarks on the repo's real manifests show
the flag is noise on npm ci with a warm cache (root: 0.90s vs 0.84s
avg; ws: 4.02s vs 4.00s avg) — npm ci does no resolution and the
content-addressed cache already serves tarballs locally. It also
carried the stale-resolution risk on the npm install fallback the
reviewer flagged. All the real win is the skip itself (0s vs ~5s+).
Tests: workspace-glob edit (desktop), literal-listed edit, and
new-workspace-under-glob all defeat the skip; verified against the
real repo's workspace config (8 manifests picked up).
Provider auto-detection (URL-based inference for Anthropic, OpenAI Codex,
and xAI endpoints) runs before credential-pool validation in AIAgent init,
but #63048 placed the pool validation before auto-detection. When the agent
is constructed with provider=None and a recognized endpoint URL, the pool
is validated against an empty provider identity and discarded, even though
auto-detection correctly resolves the provider moments later.
Fix: move the credential-pool validation block to after the URL-based
auto-detection chain. The pool is stored on the agent before
auto-detection; validation now checks the resolved provider and only
nullifies agent._credential_pool when the pool's scoped provider genuinely
doesn't match.
Regression test covers all three auto-detection paths:
- Anthropic (api.anthropic.com)
- OpenAI Codex (chatgpt.com/backend-api/codex)
- xAI (api.x.ai)
Fixes#63425.
- Nudge text now warns that repeated protocol violations will block the
task and require manual intervention, so the model understands the
consequence of ignoring the nudge.
- Kanban docs restructured to clearly separate the two defense layers:
agent-side prevention (nudge, from #64350) and dispatcher-side
recovery (bounded retry, from this PR).
- Two new integration tests verifying the nudge mentions blocking and
that the agent-side and dispatcher-side budgets are independent.
Fold in kevinb361's suggested lifecycle wording (#61817 conceded in favor of
this PR) and update the second stale site his sweep didn't cover: the
task-events table still said the dispatcher 'auto-blocks immediately instead
of retrying'. Both now describe the violation-only streak: protocol_violation
fires on every violation (its payload marker feeds the budget), below-budget
runs return the task to ready, and gave_up + auto-block happen only when the
consecutive streak reaches _PROTOCOL_VIOLATION_FAILURE_LIMIT (default 3,
per-task max_retries overriding).
Address the hermes-sweeper review of #61233: the bounded retry budget
is now a clean-exit-specific streak, not a share of the unified
consecutive_failures counter.
- detect_crashed_workers stamps a protocol_violation marker into the
violation run's metadata (via the event payload _end_run copies);
_protocol_violation_streak derives the streak from run history:
consecutive most-recent violation runs, rate_limited runs neutral
(mirroring their unified-counter treatment), any other closed run
resets it. Mixed failure kinds can neither consume nor extend the
budget.
- Below-budget violations no longer call _record_task_failure at all:
the task returns to ready with last_failure_error stamped directly
(including the corrective retry guidance wording adopted from #61817,
which build_worker_context surfaces to the retry worker) and the
unified counter is untouched, keeping the two budgets independent.
- At the bound the trip funnels through _record_task_failure with a new
keyword-only force_trip=True: the reaper has already resolved the
per-task max_retries override against the violation streak itself, so
the threshold comparison is skipped rather than double-applied.
max_retries keeps its documented top precedence in both directions:
max_retries=1 blocks on the first violation, max_retries=5 blocks on
the fifth consecutive one, unset uses the default bound of 3.
- Replace the first-violation-blocks regression test with five tests:
first occurrence retries (ready + guidance stamped + no gave_up +
unified counter untouched); streak trips exactly at the bound with
protocol_violations/protocol_violation_limit in the gave_up payload
and the auto-blocked side channel set; a prior nonzero crash does not
consume the violation budget; a non-violation failure between
violations resets the streak; max_retries precedence both directions.
All five fail against the previously reviewed diff and pass with this
follow-up. The test harness resolves hermes_cli.kanban_db fresh and
uses that single module object for the exit registry, liveness patch,
and reaper — earlier suite tests reload the module, and the old
mixed-object harness made _classify_worker_exit return unknown (the
reason the old test failed in full-suite runs on main).
Kanban suite: zero introduced failures vs upstream/main tip (62 vs 63
pre-existing environmental failures — the one no longer failing is the
old violation test this replaces; 662 passed vs 657).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
A worker that exits 0 without calling kanban_complete/kanban_block
(model stops early, transient tool wedge) tripped the failure breaker
on FIRST occurrence and the task was blocked. These are overwhelmingly
transient: with a bounded retry (limit 3, tracked via a violation
fingerprint) ~96%% of them complete on respawn. Genuine repeat
offenders still trip the breaker at the limit.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
scan_skill_commands() had two collision bugs in the same loop body:
1. Core-command collision: a skill whose normalized slug matches a core
Hermes command name or alias (e.g. "skills", "learn", "bg") would
get an auto-generated /command that shadows the core command in the
gateway dispatch path (skill map is consulted before built-in
handlers). The skill command silently overrode the core command.
2. Inter-skill slug collision: the seen_names set deduped on the raw
frontmatter name, but the command map was keyed by the normalized
slug. Two distinct names collapsing to the same slug (e.g.
"git_helper" vs "git-helper") both passed the dedup, and the second
silently clobbered the first.
Fix: add two guards in scan_skill_commands() after slug normalization:
- resolve_command(cmd_name) check skips skills colliding with any core
CommandDef (name or alias), logging a warning. Uses the existing
resolve_command() API so aliases and case variants are covered
without a separate cache. The skill remains loadable via /skill.
- cmd_key in _skill_commands check dedups on the resolved slug,
first-wins (preserving local-before-external precedence), logging
a warning naming the shadowed skill.
Combines and supersedes #31204 (@cyrkstudios), #53450 (@Gridzilla),
#50304 (@petrichor-op), and #63305 (@Vissirexa).
Co-authored-by: cyrkstudios <cyrkstudios@users.noreply.github.com>
Co-authored-by: Gridzilla <Gridzilla@users.noreply.github.com>
Co-authored-by: petrichor-op <petrichor-op@users.noreply.github.com>
Co-authored-by: Vissirexa <Vissirexa@users.noreply.github.com>
synchronous=FULL issues plain fsync(), not F_FULLFSYNC. The
F_FULLFSYNC barrier comes from checkpoint_fullfsync=1, set by the
separate _apply_macos_checkpoint_barrier(). The original docstring
conflated the two PRAGMAs.
On Darwin, the default synchronous=NORMAL only calls fsync(), which Apple
explicitly states does not guarantee data-on-platter or write-ordering.
During a WAL checkpoint race with process termination (e.g., launchd
shutdown), this can leave the main DB with half-written btree pages,
resulting in btreeInitPage error 11 corruption.
WAL mode's durability guarantee assumes the OS honors fsync barriers; macOS
does not unless we explicitly set synchronous=FULL (which issues fsync() and
F_FULLFSYNC via checkpoint_fullfsync=1).
Previously, apply_wal_with_fallback() skipped setting synchronous=FULL when
the DB was already in WAL mode, leaving connections at the unsafe
synchronous=NORMAL default. This commit adds _enforce_macos_synchronous_full()
to always enforce synchronous=FULL on macOS after any WAL activation.
Fixes#63531
Add a bounded turn-end stop guard for kanban workers. When a worker
tries to exit with finish_reason=stop without having called
kanban_complete or kanban_block, inject up to two synthetic nudges
so the conversation loop continues instead of exiting cleanly (which
the dispatcher records as protocol_violation).
Mirrors the existing verify-on-stop pattern: same ephemeral scaffolding
flag (_kanban_stop_synthetic), same role-alternation contract, same
_pending_verification_response fallback for budget exhaustion.
Disabled by default (gated on HERMES_KANBAN_TASK env var set by the
dispatcher); kill switch via HERMES_KANBAN_STOP_NUDGE=0.
Salvaged from #62262 by @mdc2122. The original branch was 272 commits
behind main with ~538 files of stale-base reversions; this salvage
applies only the 4 substantive files (agent/kanban_stop.py,
conversation_loop.py insertion, run_agent.py _EPHEMERAL_SCAFFOLDING_FLAGS,
tests/agent/test_kanban_stop.py).
Salvage of #63888. The original fix clears stale _last_content_with_tools
on substantive tool-only turns but doesn't clear _mute_post_response, which
a prior housekeeping turn may have set. This suppresses tool progress
output via _vprint until the no-tool-call branch resets it at line ~4834
— after all tools have finished executing.
Fix: also reset _mute_post_response = False when clearing stale fallback.
Added test: verify pure housekeeping turns (content + only housekeeping
tools) still set the fallback correctly — the original use case the
fallback was designed for.
Co-authored-by: liuhao1024 <sunsky.lau@gmail.com>
A cached _last_content_with_tools response from a housekeeping-only turn
could survive a later substantive tool-only turn. When the model returned
an empty response, Hermes incorrectly finalized the older housekeeping
narration instead of invoking the post-tool empty-response nudge.
Production impact: scheduled cron jobs could return early without completing
their actual work (e.g., daily report job returning a housekeeping message
instead of producing the report artifact).
Root cause: The fallback state was only updated when a turn had both
content AND tool_calls. A turn with tool_calls but empty visible content
would skip state updates entirely, leaving stale fallback state intact.
Fix: Classify tools in every tool-call turn (regardless of visible content).
When any tool is substantive (non-housekeeping), clear the older fallback state
before processing later empty responses. This prevents two-turn-old housekeeping
narration from being treated as if it belonged to the immediately preceding
substantive tool turn.
Regression test added: tests/run_agent/test_conversation_fallback_state.py
Fixes#63860
OpenAI lets ChatGPT-plan Codex users bank rate-limit reset credits, but
until now they could only be redeemed from the Codex CLI/app or the
website. This wires the same backend API into Hermes:
- /usage on the openai-codex provider now shows "You have N resets
banked - use /usage reset to activate" (parsed from the
rate_limit_reset_credits field the /usage endpoint already returns).
- New /usage reset subcommand (CLI + gateway) redeems one banked
credit via POST .../rate-limit-reset-credits/consume with a UUID
idempotency key, mirroring codex-rs backend-client semantics
(PathStyle /wham vs /api/codex, ChatGPT-Account-Id header,
reset/nothing_to_reset/no_credit/already_redeemed outcomes).
- Guard: redemption is refused while no rate-limit window is fully
exhausted, since a banked reset restores the FULL 5h + weekly
allowance and spending it early wastes it. /usage reset --force
overrides. Zero banked credits and non-codex providers are refused
with clear messages; nothing_to_reset reports the credit was NOT
spent.
- i18n: new gateway.usage.unknown_subcommand / reset_wrong_provider
keys across all 16 locales; docs updated (cli.md, messaging index).
Tested with unit tests plus a real-socket E2E against a local fake
Codex backend exercising redeem/guard/force and the /usage hint.
Cache-decorated turns (apply_anthropic_cache_control converts string
content to [{type: text, ..., cache_control}] lists — applied BEFORE the
MoA facade since the #57675 cache-cold fix) and multimodal turns
(text + image_url parts) flattened to empty strings in
_reference_messages, which only read str content. On turn 1 of a
provider:moa session with a Claude aggregator the references received a
single EMPTY user message: Anthropic-side providers 400'd ('messages: at
least one message is required') while tolerant models answered 'no user
request is present' (live incident Jul 14 2026, preset 'closed').
Fixes, in totality:
- _reference_messages: extract visible text via
agent/message_content.flatten_message_text for user/assistant/tool
turns (skips image parts, so no base64 leaks into the advisory view);
decorated and undecorated transcripts now produce a byte-identical
advisory view (advisor cache prefix stays stable).
- image-only user turns get a placeholder instead of an empty message
(Anthropic rejects empty text blocks) or a silently dropped turn
(would break user/assistant alternation).
- degenerate-case fallback flattens structured content too.
- _attach_reference_guidance: a decorated/multimodal trailing user turn
now receives the guidance as a NEW text part appended AFTER the
cache_control-marked part (cached prefix byte-stable) instead of
falling through to a second consecutive user message (strict providers
reject user/user).
- conversation_loop MoA injection: multimodal user turns get the MoA
context appended as a trailing text part instead of being dropped;
user_prompt for the one-shot path flattens content lists instead of
str()-ing them (which leaked base64 payloads into the prompt).
Live-verified on the 'closed' preset (real OpenRouter wire, 2 user
turns, tool loop): all 4 reference calls carry the full document +
rendered tool state, end on user, zero tool-role/tool_calls; advisor
cache_write 7968 then cache_read 5909+; aggregator cache_read
14880-15237 on iterations 2+.
Co-authored-by: bo.fu <bo.fu@meituan.com>
test_probe_sends_client_context_to_gemini and
test_probe_omits_gemini_client_context_for_other_providers (added in
b8eb89f5c) patch hermes_cli.models.urllib.request.urlopen, but
probe_api_models routes requests through the
_urlopen_model_catalog_request wrapper (open_credentialed_url from the
urllib_security hardening), so the mock is never invoked and
mock_urlopen.call_args is None -> TypeError. Every CI run on main and
every PR has been failing test slice 7/8 on these two tests.
Point the patches at _urlopen_model_catalog_request, the same target
every sibling test in TestProbeApiModelsUserAgent already uses.
89/89 tests in the file now pass.
Include the Hermes client name and version with Gemini inference, model and tier checks, and TTS requests. Add focused coverage for the request headers and keep the Gemini-specific context scoped to Google Gemini endpoints.
The desktop provider panel previously rendered the curated declarations
from hermes_cli/memory_providers.py: five hindsight fields, and no panel
at all for undeclared providers like honcho (OAuth connect only). The
dashboard provider-switching rework re-pointed the shared config route
at raw plugin schemas, so the desktop began dumping every internal field
(35 for hindsight) and grew a bespoke honcho panel.
Serve both surfaces from the same route: ?surface=declared returns the
curated schema (empty for undeclared providers) with the original
config-file + env-store write semantics; the dashboard keeps the raw
plugin schema unchanged. The desktop client opts into declared.
The dashboard's dedicated memory-provider UI (4b184cbe5) excluded
memory.provider from /api/config/schema server-side. Desktop's settings
page builds its field list from that schema, so the Memory Provider
dropdown silently vanished from Desktop after v0.18.1.
- web_server.py: restore memory.provider as a select in _SCHEMA_OVERRIDES,
with options built from plugins.memory discovery (was a stale hardcoded
[builtin, honcho] list before the removal)
- plugins/memory: add list_memory_provider_names() — directory-scan-only
name listing, safe at module import time (no provider imports)
- web ConfigPage: hide memory.provider client-side instead — the Plugins
page owns the dedicated provider-switching UI there
- tests: schema contract (select present, category memory, builtin
sentinel) + invariant that every discoverable provider is selectable
Self-review nits on the Thinking-widget fix:
- type ReasoningTextPart as ReasoningMessagePartComponent and read the
typed useMessagePartReasoning() directly, dropping the ad-hoc cast
(the hook already returns text/status).
- remove useRef, now unused after deleting useSmoothReveal.
- trim the autopsy comments; the PR body carries the narrative.
#63955 made Hermes survive a broken `bash -l` (Ainz's `Directory
\drivers\etc does not exist`) by falling back to non-login `bash -c`.
But a non-login shell never sources /etc/profile, so it never gets
`…\usr\bin` on PATH — and that dir holds every coreutil the file/terminal
tools shell out to (cat, mktemp, mv, wc, head, stat, chmod, mkdir, find).
Result: `write_file` returned bytes_written:0 with an EMPTY error (the
failure text went to a missing binary's stderr) and terminal commands
exited 127. The survive-broken-login-bash fix was only half-done: it
stopped crashing but silently failed every write.
Derive Git Bash's bin dirs (mingw64/bin, usr/bin, bin, …) from the
resolved bash.exe and prepend them to the subprocess PATH on Windows, in
/etc/profile precedence order so coreutils win over same-named System32
tools (find.exe, sort.exe) inside the shell. No-op off Windows and when a
login snapshot is healthy (the snapshot re-exports the full PATH inside
the shell), so this only bites on the broken-login fallback path.
Adds _git_bash_bin_dirs() (derivation, cached) + _prepend_git_bash_dirs()
(PATH merge), plus regression tests for PortableGit/MinGit layouts and
the run-env injection ordering.
Salvage of #63862. is_output_cap_error() returns False for vLLM/LM Studio
error messages that contain 'prompt contains ... input tokens' (treated as
input-overflow signal). But parse_available_output_tokens_from_error() CAN
extract a valid available_tokens from those same messages. The
compression-disabled guard only checked is_output_cap_error(), so vLLM/LM
Studio users with compression off still got a terminal failure instead of
the max-tokens retry.
Fix: also exempt when parse_available_output_tokens_from_error() returns a
value — that function determines whether the retry path can actually handle
the error, so it's the right predicate for the exemption.
Added test: verify vLLM-format error with compression_disabled=False still
triggers the max-tokens retry path.
Co-authored-by: dmabry <dmabry@users.noreply.github.com>
The branch computed safe_out from estimate_messages_tokens_rough(messages),
but the provider rejected the larger api_messages request (system prompt,
injected context, tool schemas). When API-only content is large, safe_out
could far exceed the provider's available_tokens.
Compute safe_out from estimate_request_tokens_rough(api_messages, tools=...)
and keep provider available_out as an upper bound. Do not alter context_length
or trigger compression for output-cap errors.
Add production-path run_conversation tests that assert the retry API call's
max_tokens, including a case where a large system prompt makes messages-only
estimation undercount the real request.
Fixes#55546
The retry loop computed safe_out from the error's available_tokens,
which reflected the *previous* request. Between retries the agent
appends tool results and error text, so the real input token count
grows. Deriving safe_out from the stale budget meant every retry
still exceeded the context ceiling by 1+ tokens, burning through the
3-attempt limit.
Compute safe_out from estimate_messages_tokens_rough(messages) so
the cap tracks the growing input on each retry attempt.
`_module_registers_tools()` reads each `tools/*.py` file and fully
AST-parses it to check for a top-level `registry.register()` call.
90 files are scanned on every process start — but only 32 actually
register tools.
Add a cheap text prefilter: after reading the file (which we need to
do anyway for AST), check that both `"registry"` and `"register"`
appear in the source before calling `ast.parse`. A file with a
top-level `registry.register()` call must contain both strings, so
this is a perfect superset — zero false negatives. 50 of 90 files
skip the AST parse entirely.
The `source=` parameter is not threaded through `discover_builtin_tools`;
the prefilter lives entirely inside `_module_registers_tools`, keeping
the public API unchanged.
Benchmark (median of 10 runs, scanning 90 files):
before (read + ast.parse all): 305.9ms
after (text prefilter + ast): 187.8ms
speedup: 1.6x (118ms saved)
Identical module set: 32 modules, same names, same order.
The _PROVIDER_PREFIXES frozenset in agent/model_metadata.py is static
and does not auto-extend from ProviderProfile. Removing deepinfra and
deep-infra from it broke provider:model prefix stripping for DeepInfra.
Preserve one durable staged input across terminal close and the worker's early turn flush, without duplicating resumed transcripts or creating a session with a null prompt. Fixes#63766.
Retain a distinct CLI history baseline during the signal window before a turn's normal persistence flush. When CLI history aliases the live agent list, use marker-only persistence so a genuinely unflushed tail is written.
Salvage of #63935. The original fix read HERMES_CRON_SESSION_DB_TIMEOUT
from a bare env var, but AGENTS.md requires non-secret behavioral
settings to live in config.yaml with an env var bridge only for
backward compatibility.
Changes:
- Add cron.session_db_timeout_seconds to DEFAULT_CONFIG (default 10s)
- Resolution order: HERMES_CRON_SESSION_DB_TIMEOUT env override →
cron.session_db_timeout_seconds in config.yaml → 10s default
(mirrors the existing script_timeout_seconds pattern)
- 0 = unlimited (opt-in for debugging, skips the bound)
- Strengthen test: assert the warning is logged on invalid env value
(caplog was taken but never asserted)
- Add test: verify config.yaml resolution path works end-to-end
Co-authored-by: LoicHmh <26006141+LoicHmh@users.noreply.github.com>
run_job() constructs SessionDB() synchronously with no timeout of its
own, unlike the agent's run_conversation call further down, which is
already bounded by HERMES_CRON_TIMEOUT. A wedged sqlite3.connect (e.g.
a stale flock from a crashed sibling process) hangs this call
indefinitely.
That hang is invisible to every existing cron safeguard because it
happens before _submit_with_guard's future exists: the finally block
that discards the job ID from _running_job_ids never runs. The job
stays wedged "running" — every later tick logs "already running —
skipping" — until the whole gateway process is restarted.
Observed in production: a cron job's worker thread was confirmed via
a live py-spy thread dump to be parked inside SessionDB.__init__'s
sqlite3.connect for 3+ days, silently skipping every scheduled fire
in between across a gateway process that otherwise stayed healthy.
Bound the SessionDB() construction with its own timeout
(HERMES_CRON_SESSION_DB_TIMEOUT, default 10s), following the same
bounded-thread-pool pattern already used elsewhere in this file (the
delivery retry path, and the agent inactivity watchdog just below).
On timeout, log at ERROR and proceed with session_db=None instead of
degrading silently to debug level, since an actual hang here is a new
condition worth surfacing.
Adds tests/cron/test_sessiondb_init_hang.py, including an end-to-end
regression proving the dispatch guard is released and a subsequent
tick can fire the same job again after a simulated hang.
The Thinking disclosure rendered blank for every reasoning-emitting model
(Fable, DeepSeek, GPT-5.5, ...). Two causes:
1. ReasoningTextPart read a `text` prop that assistant-ui never populates —
reasoning parts arrive via context, same as text parts — so it always got
an empty string. Read the text via useMessagePartReasoning() instead,
mirroring how MarkdownText uses useMessagePartText().
2. The reasoning-only SmoothStreamingText / useSmoothReveal layer stalled at
revealed="": the reasoning part stays isRunning for the whole message while
the answer streams and thrashes re-renders, so the char-reveal never
advanced past 0. Render reasoning through the same DeferStreamingText →
surface path the assistant answer uses, and drop the dead smoothing code.
Baseline bars were positioned at their absolute timeline offset, which
included the baseline run's own wait time — making duration comparisons
hard since the bars were visually offset. Now each baseline bar starts
at the same left position as its corresponding current job, so the two
bars directly overlap for at-a-glance duration comparison.
Also removed the now-unused bl_t0 / bl_max / bl_jobs_timed variables
that tracked the baseline timeline position.
Add wait time computation: for each job, wait_s = started_at - max(completed_at
of all jobs that finished before it started). This is a timestamp heuristic
(no workflow YAML dependency parse needed) that's accurate for pipeline-shaped
CI where the critical path is linear at each stage.
Shows up in:
- Job table: new Wait + Δ Wait columns
- Gantt chart: hatched bar segment before the run bar
- Step details: '(wait Xs)' annotation in the summary line
- Markdown summary: Total wait row with delta vs baseline
- Stats: total_wait / bl_total_wait in compute_stats
Also completes the skipped-jobs UI:
- Skipped stat card in stats cards
- Skipped row in markdown summary table
Backward compatible: old cached baselines without wait_s annotate on load.
Skipped jobs (conclusion == 'skipped') have null/zero-duration timestamps
that polluted every downstream computation: they counted as 'unchanged'
(0 vs 0) in faster/slower tallies, showed meaningless '0.0s (0%)' deltas
in the job table, rendered phantom gantt bars, and inflated wall/compute
totals.
Add is_skipped() helper and apply it consistently:
- compute_stats: exclude from wall/compute + faster/slower/unchanged;
add 'skipped' and 'bl_skipped' counts
- _gantt_bars: filter from current bars, baseline bars, and axis calc
- _job_table: show 'skipped' label instead of durations/deltas
- _step_details: skip entirely (no meaningful step data)
- _regressions: exclude from both current and baseline sides
alelpoan's fix (emit hermes:zoom:changed after restore) is correct, but the
bug's root is duplication: setAndPersistZoomLevel and restorePersistedZoomLevel
each independently did setZoomLevel + send, and restore forgot the send.
Collapse both (and the lifecycle re-assert) into a single applyZoomLevel()
helper in zoom.ts that always applies-then-notifies — the regression can't
recur by omitting a send. Replace the source-grep test (which broke on main:
the sibling source-assertion pet test it copied was refactored to a behavioral
one, dropping the fs/path imports it relied on) with behavioral coverage of the
funnel, matching zoom.ts's "unit-testable without booting a BrowserWindow"
convention.
Co-authored-by: alelpoan <alelpoan@proton.me>
classifyNativeBinary only checked big-endian Mach-O/Fat magic bytes
(feedfacf, feedface, cafebabe). Real Darwin .node files from node-pty
prebuilds are stored little-endian on disk (cffaedfe = MH_CIGAM_64),
so every Darwin prebuild classified as null, and validateStagedBinaries
threw a platform mismatch on macOS — breaking npm run check for every
macOS contributor. CI didn't catch it because runners are Linux (ELF
path was correct) and tests only planted big-endian fake headers.
Add recognition for all six Mach-O/Fat byte orderings:
- MH_CIGAM (cefaedfe) — LE 32-bit
- MH_CIGAM_64 (cffaedfe) — LE 64-bit [the one real prebuilds use]
- FAT_CIGAM (bebafeca) — LE universal
Update makeFakeNode to write LE CIGAM_64 bytes for the darwin fixture
(matching real on-disk format) and add regression tests for all new
magic forms.
The set-matrix step wrote the npm workspace query result directly to
$GITHUB_OUTPUT. If discovery ever produced [], the matrix would expand
to zero check jobs, leaving the reusable workflow green without running
any JS/TS checks.
Now the step validates the result is a non-empty array before emitting
it, and exits 1 with a GitHub annotation if it's empty or jq failed.
stageNodePty received electron-builder's { platform, arch } but unconditionally
copied host build/Release, staging a host binary (e.g. macOS Mach-O) for a
foreign target (e.g. linux-arm64). The fallback rebuild also didn't pass the
target arch to electron-rebuild.
Now:
- build/Release is only staged when target platform+arch match the host
- cross-platform targets with no matching prebuild fail closed
- same-platform different-arch rebuild passes --arch to electron-rebuild
- post-staging validation reads .node magic bytes (ELF/Mach-O/PE) and rejects
any binary whose platform doesn't match the target
Adds stageNodePtyInto() (testable core) and classifyNativeBinary() (pure),
plus 11 regression tests covering the cross-target scenarios.
Move setGatewayState + rerender calls inside act() blocks and make the
synchronous soft-switch test async so all state updates are wrapped.
Eliminates the last 4 act() warnings (44 → 0).
Extracted zoomWiringForWindowKind() + ZOOM_WINDOW_CONFIG into zoom.ts so
the pet-overlay-
opts-out / chat-windows-keep-zoom contract is tested via the pure
config,
not by reading source. Callers in main.ts now use
zoomWiringForWindowKind()
instead of inline { zoom: false } / default { zoom: true }.
When neither a prebuild nor a compiled build/Release/*.node is found for
the target platform-arch, stage-native-deps.mjs now runs
electron-rebuild -f -w node-pty to compile one from source before
re-copying build/Release into the staged dist.
This makes the staging script self-sufficient — it always produces a
working native binary dir regardless of whether npm ci --ignore-scripts
skipped postinstall or whether node-pty publishes prebuilds for the
target (e.g. linux-x64 has no prebuild).
Add vitest.setup.ts with IS_REACT_ACT_ENVIRONMENT=true + auto-cleanup,
and wrap render()/fireEvent() calls in act() across 8 test files:
- provider-config-panel.test.tsx: wrap renderPanel + fireEvent in act
- providers-settings.test.tsx: wrap renderProvidersSettings + fireEvent
in act
- use-prompt-actions/index.test.tsx: add actRender helper, wrap all 38
render
calls, wrap Harness handle methods (submitText/cancelRun/steerPrompt/
restoreToMessage) in act at the onReady callback level
- attachments.test.tsx: make renderWithI18n async + wrap in act
- skills/index.test.tsx: make renderSkills async + wrap fireEvent in act
- messaging/index.test.tsx: make renderMessaging async + wrap fireEvent
in act
- gateway-connecting-overlay.test.tsx: wrap all render/rerender in act
- preview-pane.test.tsx: wrap render calls in act, make tests async
Reduces act warnings from 44 to 4 (remaining are fake-timer + pre-render
store mutation edge cases). All 146 test files / 1180 tests still pass.
textInputCursorSourceOfTruth.test.ts read textInput.tsx as text and
regexed it to check that cursorLayout() is called with curRef.current
(not the stale cur React state), and that the fast-echo backspace/append
stdout writes are paired with noteCursorAdvance calls.
Extract three pure functions from textInput.tsx:
- resolveCursorLayout(display, cur, curRefCurrent, columns): wraps
cursorLayout(display, curRefCurrent, columns), making the
curRef.current-over-cur choice a directly testable pure call instead of
a regex match on the render-site call expression.
- fastBackspaceEffect(current, cursor) / fastAppendEffect(current, cursor,
text): return a single object bundling {newValue, newCursor, write,
advanceDelta} for each fast-echo path. Bundling the stdout write and the
noteCursorAdvance delta into one return value makes the pairing
impossible to silently drift apart (a caller can't get without
), instead of relying on the two call sites appearing near
each other in source text.
textInput.tsx's render site and backspace/append handlers now call these
helpers directly, preserving exact existing behavior (the same '\b \b'
write sequence, noteCursorAdvance(-1)/noteCursorAdvance(text.length) calls).
textInputCursorSourceOfTruth.test.ts imports and calls the three pure
functions directly with a deliberately stale cur vs a fresh curRefCurrent
to reconstruct the exact regression scenario, and asserts the bundled
effect objects -- no readFileSync, no regex against textInput.tsx's
source text. Full ui-tui suite (107 files, 1117 tests) still green.
oauth-session-request.test.ts regexed main.ts source text (extracting the
fetchJsonViaOauthSession function body and matching regexes against it) to
check Electron net.request doesn't set the forbidden Content-Length header
and does call request.write(body).
That behavior is already covered for real by oauth-net-request.test.ts,
which imports the actual serializeJsonBody/setJsonRequestHeaders helpers
from oauth-net-request.ts and asserts on a mock request object's setHeader
calls -- it only needed its relative import corrected to include the .ts
extension (Node's ESM loader doesn't resolve extensionless relative
specifiers). Removed the now-fully-superseded oauth-session-request.test.ts
and its dangling package.json wiring.
profile-delete-respawn.test.ts regexed main.ts source text to check that
prepareProfileDeleteRequest returns the torn-down profile name, and that
the hermes:api ipcMain handler captures that return value and routes to
the primary backend instead of respawning a pool backend for the just-
deleted profile.
Extract the pure decision logic into profile-delete-routing.ts (no
Electron import):
- profileNameFromDeleteRequest(request): parses a DELETE
/api/profiles/<name> path, moved verbatim (already pure).
- decideProfileDeleteAction(profile, deps): the branch decision (noop /
teardown-primary / teardown-pool) and the profile name to return,
parameterized over isDefaultProfile/isValidProfileName/primaryProfileKey.
- resolveRouteProfile(tornDownProfile, profile): the
routing ternary from the hermes:api
handler.
prepareProfileDeleteRequest in main.ts now calls decideProfileDeleteAction
for the decision and only performs the async side effects (teardown +
writeActiveDesktopProfile) the decision calls for.
profile-delete-routing.test.ts replaces profile-delete-respawn.test.ts
(which was never wired into test:desktop:platforms), importing the pure
functions directly and asserting real return values across every branch
-- no readFileSync, no regex-on-source. Wired the new test file into
test:desktop:platforms in package.json.
windows-hermes-resolution.test.ts regexed main.ts source text to check
three Windows resolution bugs that caused desktop reinstall loops:
1. findOnPath()'s PATHEXT extension order (must try real extensions before
the empty one, or an extensionless Git-Bash hermes shim shadows
hermes.cmd/.exe).
2. handOffWindowsBootstrapRecovery()'s --update vs --repair choice (must
gate on any real-install signal, not just the hermes.exe shim).
3. unwrapWindowsVenvHermesCommand()'s probe-before-trust behavior (must
canImportHermesCli() before returning a venv python, or a broken venv
gets re-selected forever).
Extract all three into pure, dependency-injected functions in
windows-hermes-path.ts (no Electron import): buildPathExtCandidates(),
chooseUpdaterArgs(), resolveVenvHermesCommand(). main.ts's
findOnPath/handOffWindowsBootstrapRecovery/unwrapWindowsVenvHermesCommand
now call these with their existing helpers (fileExists, canImportHermesCli,
getVenvPython, etc.) passed through as deps.
windows-hermes-path.test.ts replaces windows-hermes-resolution.test.ts,
importing the pure functions directly and asserting real return values
with fake/injected dependencies (fake venvs, fake probes) -- no readFileSync,
no regex-on-source.
windows-child-process.test.ts regexed main.ts and bootstrap-runner.ts
source text to check that spawn/execFileSync call sites wrapped their
options with hiddenWindowsChildOptions(), and that backend teardown
chose the right kill strategy.
Extract both into dependency-free sibling modules:
- windows-child-options.ts: hiddenWindowsChildOptions(options,
isWindows)
now takes isWindows as an injectable param (defaults to the real
platform check). main.ts and bootstrap-runner.ts both import the same
implementation instead of each defining their own copy.
- backend-child.ts: stopBackendChild(child, deps) with
forceKillProcessTree
and isWindows injected, so the SIGTERM-vs-tree-kill branching is
directly
testable with a fake child + a spy.
windows-child-options.test.ts replaces windows-child-process.test.ts,
calling the real functions with fake spawn/execFileSync-shaped objects
and asserting on the actual returned options / kill call.
Same fix as the base-env bootstrap tests: init_session's post-failure
non-login probe was overwriting captured["script"] with `true`. Capture
the first (bootstrap) _run_bash call via setdefault.
init_session now fires a follow-up non-login `bash -c true` probe when the
login bootstrap fails, so the two bootstrap-script assertions must capture
the first call only (setdefault), not the probe that overwrote it.
Commit 62a76bd3d (feat: make smart approvals the default, #62661)
changed approvals.mode default from "manual" to "smart". The
TestBlockingApprovalE2E tests did not patch the approval mode, so
check_all_command_guards routed through _smart_approve() first —
calling call_llm() which tried all auxiliary providers, failed (no
API keys in test env), and returned "escalate" before falling through
to the gateway blocking path. The LLM failure cascade took longer
than the test's 2.5s wait window (50 × 0.05s), so the notify
callback had not fired yet when assert len(notified) == 1 ran.
Force {"mode": "manual"} via patch in setup_method/teardown_method,
matching the pattern already used by test_blocking_approval_uses_
canonical_timeout in the same class.
#63621 fixed path quoting, but Ainz's Git for Windows still dies on
`bash -l` itself (`Directory \drivers\etc`). Hermes then fell back to
bash -l *per command*, so every write_file/terminal call failed the same way.
After a failed login snapshot, probe non-login bash -c; if it works, skip
-l for the session. Also skip a stale HERMES_GIT_BASH_PATH that fails a
noprofile probe in favor of %LOCALAPPDATA%\hermes\git portable bash.
Live deltas from session A were attaching to session B after New Session
when events arrived without session_id — fallback used the newly focused
activeSessionId (#47709).
Pin unscoped stream events to the session that received message.start
(#48281). Also reset RAF-pending view staging on new/resume/create so a
stale background flush cannot repaint over the switched chat (#47743).
Co-authored-by: Ray <rayjun0412@gmail.com>
Co-authored-by: zapabob <1920071390@campus.ouj.ac.jp>
When the base is an origin/… ref, fetch just that branch so the
local tracking ref is fresh before `git worktree add -b new origin/main`.
Fetch failures (offline / no remote) are silently ignored — git uses
whatever local ref exists, or raises a clear error if it's missing.
The sidebar "New worktree" button branched off whatever HEAD you were
on — now a filterable Popover+Command combobox lets you pick any local
or remote-tracking branch as the base, defaulting to origin/HEAD.
Backend listBaseBranches() queries refs/heads + refs/remotes via
for-each-ref, flags origin/HEAD (falling back to the local default for
no-remote repos). Mirrored on the Python REST API side
(base_branch_list + /api/git/base-branches route).
Splits the single broad `except Exception` in the compression-lock
acquire path into two handlers: AttributeError/TypeError (version skew —
the lock method is missing, or predates the `ttl_seconds=` kwarg) still
fails OPEN as before, since that's known-safe to proceed without a lock.
Any other exception now fails CLOSED (skips compression this cycle)
instead of treating every failure as "no lock subsystem present" and
letting a second compressor run concurrently and fork the session.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Reserve API requests before their first await, include every drain-owned work type in runtime status, and pause local cron dispatch while a gateway drain is active.
Closes#63529
Root cause: GatewayRunner._drain_active_agents only waited on
_running_agents + cron in-flight counts. Desktop/API sessions are
tracked solely inside APIServerAdapter (_inflight_agent_runs +
_active_run_agents), so stop/restart logged active_at_start=0 and
systemd SIGKILL'd mid-tool work.
Fix: APIServerAdapter.active_agent_work_count() plus
GatewayRunner._active_api_run_count() folded into the drain wait,
status updates, timeout result, and shutdown logs — same pattern as
_active_cron_job_count for #60432.
Verification: pytest tests/gateway/test_api_server_active_work_drain.py
tests/gateway/test_cron_active_work_drain.py -q → 19 passed
Normalize Anthropic setup-token metadata for every PooledCredential construction path, persist corrected manual entries, heal legacy rows on load without copying global fallback credentials into profiles, and map the contributor email for release attribution.
Manually-added Anthropic setup-tokens defaulted to api_key and were sent
via x-api-key, which Anthropic rejects -> 429. Infer OAuth from the
sk-ant-oat prefix in from_dict so all ingest paths agree with
_is_oauth_token(). Fixes#63737.
Add fallback only updates local editor state; complete pairs are filtered
before onChange. The post-#7b5ba205 resync effect then saw the unchanged
persisted chain and wiped the draft — button looked dead.
Ignore value updates that match the last chain we emitted; still resync
on real external changes (profile/config reload).
Co-authored-by: HexLab98 <liruixinch@outlook.com>
Reject metadata that would make session queries fail, bound import work, and detach cyclic lineage links. Guard lineage traversal against pre-existing corrupt cycles.
_custom_provider_model_matches() only compared the session model
against the entry's single 'model' field. A custom provider declaring
a multi-model catalog (providers.<name>.models mapping / models list)
whose default model differed from the session model silently failed to
match — dropping the entry's extra_body entirely. Real impact: an
OpenAI custom provider pinning service_tier=flex via extra_body ran
every request at STANDARD tier (~2.3x billing) with zero signal.
- Model matching now accepts the session model when it appears in the
entry's models catalog (dict keys or list), case-insensitive;
single-model 'model' field behavior unchanged; entries with neither
still match everything.
- Usage report ('hermes -z --usage-file') now carries service_tier
(the tier requested via request_overrides.extra_body) so batch
pipelines can audit the billed tier per run.
Validation: 8 new tests; live E2E via real 'hermes -p sweeper -z'
with httpx-level wire capture — service_tier=flex present in the
outgoing /v1/responses body and in the usage report.
_default_spawn sets HERMES_KANBAN_GOAL_MODE=1 but launched 'chat -q' without
-Q; _run_kanban_goal_loop_q only executes in the quiet single-query branch,
so goal-mode never ran for dispatcher-spawned workers — they got one turn,
printed text, exited rc=0, and tripped the protocol-violation circuit
breaker (2026-06-09, cards t_d9cbe312 et al). Root-cause report + upstream
issue draft in kanban workspace t_720c5c60.
#63113 rewrote native drive paths in ShellFileOperations, but init_session
/_wrap_command still embedded C:/... hermes-snap paths from get_temp_dir.
MSYS arg-converts those during bash -l and surfaces Directory \drivers\etc
— including for relative write_file targets, since the wrapper is the fault.
Add _bash_safe_path, override BaseEnvironment._quote_shell_path on
LocalEnvironment (no base→local import), and normalize mixed /c/Users\...
paths in file ops.
Co-authored-by: xxxigm <tuancanhnguyen706@gmail.com>
The #54527 context pin (7acaff5ef) snapshots the selected stored session
and route token at submit entry and aborts when either changes mid-flight.
But a NEW chat's create pipeline legitimately moves both: on success,
createBackendSessionForSend re-homes selection and navigates to the chat
it just minted. Judged against the pre-create draft baseline that read as
a user switch, so every first send of a new chat aborted before
prompt.submit — message dropped, no DB row persisted (row creation is
lazy, server-side in prompt.submit), and the window stranded on a route
whose REST reads 404 "Session not found" forever.
Fix: after a successful create, verify no one re-homed during create's
post-commit await via the active-session ref (a non-null return
guarantees create set it; every switch path retargets it synchronously),
then re-pin the drift baseline to the created chat. A mid-create switch
still aborts through create's own null return, or through the active-ref
check for the post-commit window. Re-pinning also restores the correct
stored-id association for the optimistic-message state updates, which the
pinned pre-create null had degraded.
Tests: red-first regression for the new-chat send, an abort case for a
switch landing in create's post-commit window, and the sleep/wake
new-chat stub made faithful to the real create (it sets the active ref
before returning — the inert stub is what let this ship green).
The web UI CONFIG_SCHEMA showed ['ask', 'yolo', 'deny'] for the
approvals.mode select field. These don't match any real config values
and 'smart' mode was entirely unreachable from the dashboard.
Correct the options to ['manual', 'smart', 'off'] which match the
values defined and documented in hermes_cli/config.py.
Adds a regression test to TestBuildSchemaFromConfig to pin the correct
option names and guard against future drift.
Fixes#31925
Asserts each {provider, model} entry renders as its own row (the bug
produced "[object Object]"), that removing a row emits the remaining
entries, that adding a blank row never persists a partial pair, and the
empty-state hint.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Settings → Model rendered `fallback_providers` (a list of `{provider,
model}` objects) through the generic `list` config field, which does
`value.join(', ')` and stringified each entry to `[object Object],
[object Object]`.
Add a dedicated provider+model row editor (add/remove), sourced from the
same `getGlobalModelOptions()` the composer picker uses, that reads and
writes the `{provider, model}` chain. Half-filled rows are kept in local
state so the config autosave never persists a partial entry, and an
out-of-catalog model stays selectable so existing custom entries render.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Preserve the root cause and precedence direction from #43538 while applying the merge before truncation and covering all declared model shapes.
Co-authored-by: liuhao1024 <sunsky.lau@gmail.com>
list_authenticated_providers() emits picker rows for every slug in
PROVIDER_TO_MODELS_DEV that has any credential env-var set. Several of
those slugs (notably 'mistral') have no PROVIDER_REGISTRY entry, so
resolve_provider() rejects them as 'Unknown provider' once the user
selects a model — leaving the picker showing rows that cannot actually
be selected.
Add a resolve-gate in section 1: if PROVIDER_REGISTRY.get(hermes_id)
is None, skip the slug. The picker now only lists providers that can
actually be switched to at runtime.
This automatically resolves the duplicate-Mistral dedup symptom too:
once the broken-from-models.dev row is filtered, the conflict between
PROVIDER_TO_MODELS_DEV['mistral'] and a custom_providers 'Mistral' row
is moot.
Composes with #50289 (which promotes mistral to first-class via the
provider-plugin path): when that lands, PROVIDER_REGISTRY gains a
'mistral' entry and the gate becomes a no-op for it. No conflict.
Tests (regression suite):
- tests/hermes_cli/test_model_switch_filter_unresolved.py (new, 4 tests):
Picker excludes 'mistral' when MISTRAL_API_KEY is set; 'deepseek' and
'xai' (PROVIDER_REGISTRY-backed) still appear; 'mistral' stays
excluded when no key is set. Confirmed by reverting the fix and
seeing the test fail with 'mistral leaked into /model picker'.
Cross-checked against the existing 51 test_model_switch_* and
test_custom_provider_* cases — 55/55 PASS, no regressions.
MoA was internally inconsistent: preset-level ops (set default / add /
delete) persisted on click, but reference-model and aggregator slot edits
sat behind a manual Save button. Debounce-persist slot/aggregator edits
like the rest of settings and drop the redundant button, so MoA is
uniformly autosave.
An aggregator whose pooled credentials are all exhausted/dead still counted as
an authenticated provider during no-provider /model resolution. It then won the
model-name match, was set as the sticky session provider, and poisoned every
later switch with "empty API key" errors while still routing through the dead
aggregator.
list_authenticated_providers now requires a pool to have at least one available
entry (has_available, not has_credentials / bare key presence) at all three
credential-pool gates. Simple token-style entries that don't parse into
exhaustion-tracked entries keep the prior behaviour, so providers whose creds
live only in the auth-store credential_pool still appear.
Fixes#45759
ShellFileOperations builds bash commands (wc/head/sed/cat/tee ...) with the
target path as an argument. On a Windows/Git-Bash host a native `C:\...` path
has its backslashes eaten by bash (and mangled by the msys runtime even when
single-quoted) — the "Directory \drivers\etc does not exist; exiting — update
your msys package" class of failures. Rewrite a native drive path to forward
slashes in `_escape_shell_arg`, reusing the env layer's `_windows_to_msys_path`.
Both `C:/...` and `/c/...` fix the backslash bug (the MSYS coreutils resolve
either via the POSIX API). We emit `/c/...` purely for consistency: it's the
same form `_windows_to_msys_path` already produces for the terminal `cd`
(LocalEnvironment._quote_cwd_for_cd), so shell file ops and `cd` share one
helper and one path form.
Scoped from #55481, which also patched BaseEnvironment._quote_cwd_for_cd — but
LocalEnvironment already overrides that through `_windows_to_msys_path`, so on a
real Windows host the base branch never ran (the cwd is already `/c/...`).
Co-authored-by: konsisumer <der@konsi.org>
Add shared translators in hermes_constants (Windows drive → /mnt, `\\wsl(.localhost|$)\`
UNC → POSIX, gated on is_wsl) and apply them at the gateway session-cwd boundary
so a Windows-host UI can hand the WSL backend a path it can actually chdir into.
De-dups the ACP adapter's private `_win_path_to_wsl` onto the shared helper and
extends it to the UNC spelling.
Co-authored-by: Rage Lopez <VrtxOmega@pm.me>
When the desktop UI runs on Windows and the gateway runs in WSL, a WSL/POSIX
cwd isn't openable/readable from the Windows host. Add wsl-path-bridge.ts to
translate the Windows-side direction only:
- native folder dialog defaultPath: `/home/...` → `\\wsl.localhost\<distro>\...`
- fs read path: WSL cwd → its UNC / `C:\` drive form
Distro detection reads `wsl.exe -l -q` with `WSL_UTF8=1` and strips stray NUL
bytes, since older wsl.exe emits UTF-16LE (microsoft/WSL#4607) — the original
utf8 read returned a garbled distro name. UNC uses `\\wsl.localhost\` with a
`\\wsl$\` fallback for older Windows. The reverse (any path → POSIX) is handled
once gateway-side, so the picker result needs no desktop translation.
Co-authored-by: Rage Lopez <VrtxOmega@pm.me>
Capture durable Desktop engineering principles from recent sessions —
state by authority, workspace-switch shapes, resolver ladders, optimistic
UI — and point root AGENTS.md at the scoped guide with current filenames.
Answered clarifies were collapsing into a generic tool row, hiding the
choice. Settle into a Q&A panel instead, and route freeform input through
the shared Textarea chrome.
Resuming a session cd's back into its recorded working directory, so it resumes
in the repo it belonged to. `--no-restore-cwd` opts out; skipped under
--worktree (that path owns its dir); best-effort — a missing dir warns and stays
put rather than failing the resume.
Co-authored-by: Cary Palmer <palmer@dugoutfantasy.com>
`hermes sessions list --workspace <needle>` filters to one workspace (git repo
root or project dir, matched by path substring or basename) and adds a
Workspace column. The column only appears once at least one listed session
carries a workspace, so all-unbound listings render exactly as before.
Co-authored-by: Cary Palmer <palmer@dugoutfantasy.com>
A session's coarse workspace identity: its git repo root when known, else its
cwd (branch excluded, so switching branches doesn't fragment history). Pure
helper over fields sessions already record — no new columns, no git shelling.
Co-authored-by: Cary Palmer <palmer@dugoutfantasy.com>
Align the workspace status-bar dropdown with the rest of the status bar: drop
the per-item icons (they mixed lucide size-4 with a Codicon 1rem glyph and were
the only status-bar menu carrying item icons), leaving text-only items on the
shared DropdownMenuItem primitive with default typography. The status-bar
trigger keeps its FolderOpen glyph, consistent with sibling items.
Also map true@supersynergy.de → Supersynergy in AUTHOR_MAP.
Squashed salvage of #45744 (@harjothkhara), rebased onto current main and
resolved against #58241 (which swapped the new-session cwd fallback to the
project-aware resolveNewSessionCwd).
An explicitly clicked sidebar workspace stays authoritative until session.create:
a one-shot $newChatWorkspaceTarget (null → detached, string → that folder) plus a
generation counter so a stale async `config.get project` normalization can't
overwrite a newer draft target. The start-workspace-session action is extracted
out of desktop-controller.tsx into a testable workspace-session-target module.
Integrated with #58241: the no-explicit-target branch now falls through to the
project-aware resolveNewSessionCwd() instead of the old workspaceCwdForNewSession.
Co-authored-by: harjoth <harjoth.khara@gmail.com>
Addresses @teknium1's review of #61950:
- The desktop live overlay (workspace-groups.ts) matched cwd membership
case-sensitively, so a fresh mixed-case/separator Windows session missed
its explicit/auto project until the next backend tree refresh. Mirror the
backend identity (isWindowsPath/comparisonSegments/pathKey) in isPathUnder,
liveSessionProjectId, and overlayRepoLanes lane matching. Comparison-only —
emitted ids/labels keep their spelling. POSIX stays case-sensitive.
- Backend _is_windows_path missed root-relative `\wsl.localhost\...` (single
leading backslash), leaving that historical spelling case-sensitive. Classify
any backslash-rooted path as Windows.
Tests: WSL-spelling collapse + explicit-project precedence (project_tree),
Windows/WSL live-overlay membership + POSIX case-sensitivity (workspace-groups).
Sign out of the dedicated OAuth partition before opening the login window so a
stale gateway/identity-provider cookie can't silently bounce an expired session
straight back into failure. Relabel the action "Sign out & sign in" and spell
out the sign-out step in the hint.
Co-authored-by: Tony Antunez <57689194+smtony@users.noreply.github.com>
Add isRemoteReauthError so an auth-shaped boot error counts as a remote-reauth
failure even when the session indicator still reads connected (a stale refresh
cookie / failed ws-ticket mint). Wire the boot error into the overlay's reauth
check so those sessions route to Sign in instead of the local-only recovery
buttons.
Co-authored-by: Tony Antunez <57689194+smtony@users.noreply.github.com>
A remote/VPS backend that failed to boot trapped the user on the recovery
screen — Retry/Repair/Use-local only target the local backend, so the only fix
was hand-editing connection.json. Add an in-place "Gateway settings" view (the
real GatewaySettings panel embedded via `embedded`, lazy-loaded) reached from
the recovery card, and shape the recovery actions by failure kind: Sign in for a
lapsed remote session, Gateway settings for any other remote failure (Retry
drops to secondary; Repair is dropped — it can't revive a remote), Retry for a
local backend. Use-local is scoped to remote failures.
Add an `embedded` flag to GatewaySettings (and a `bare` variant to
SettingsContent) that drops the page title/intro, Diagnostics row,
"Save for next restart", and the page gutters — so the same panel can be reused
inside a tighter surface without a second connection form to maintain. No change
to the standalone Settings → Gateway page (defaults off).
Factor the "remote/cloud with a URL" check out of isRemoteReauthFailure into a
shared isRemoteConfig helper so the boot-failure overlay can tell any remote
failure apart from a local one.
Remove the DEV-gated "Dev · soft switch" ListRow and its previewGatewaySwitch
helper. It was a temporary review affordance for exercising the soft-switch
reconnect; dead-stripped from production, but it doesn't belong in the tree.
wipeSessionListsForGatewaySwitch (the real path) and $gatewaySwitching stay.
A block-level label child (e.g. `flex`) collapses TooltipContent's inline
`box-decoration-clone` wrapper, so Radix measures a zero-size chip and parks
an empty black rectangle in the panel corner instead of by the trigger
(#62022). The terminal rail's hotkey labels and the preview row's two-line
label both hit this.
Harden the shared wrapper (`[&>*]:!inline-flex`) so any call site's direct
child renders inline-flex, add a reusable `TipHintLabel` for the common
text+hotkey label, and keep the preview row's label explicitly inline-flex.
Salvages #62139 (shared-component hardening + TipHintLabel) and #62073
(inline-flex call-site fixes + rail/preview coverage).
Co-authored-by: alelpoan <alelpoan@proton.me>
Co-authored-by: zapabob <1920071390@campus.ouj.ac.jp>
A reopened tab restarted the shell in its original launch dir, so the fresh
prompt showed the wrong folder after a prior `cd` (the issue's "separate
thing" note). Track the shell's working directory and restart the PTY there.
Two independent signals feed a persisted per-tab restoreCwd:
- a main-side PTY cwd probe (shell-agnostic; /proc on Linux, lsof on macOS;
Windows has no cheap per-process query so it falls back to the launch dir)
- cwd-reporting OSC sequences parsed in the renderer (OSC 7 file URIs, OSC 9;9
ConEmu/Windows-Terminal paths) for shells configured to emit them
On relaunch the fresh shell boots in restoreCwd, falling back to the launch
cwd (then home) when it no longer exists.
cleanReviveSnapshot only dropped the trailing prompt when a blank separator
sat above it (starship add_newline), so shells that print the prompt with no
preceding blank line — default PowerShell (PS C:\..>), bash user@host:~$ —
kept the idle prompt in the saved buffer and showed a duplicate under the
fresh boot prompt on every relaunch of an *active* session.
An interactive shell always reprints its prompt after a command, so the tail
of an idle buffer is the prompt, never history. Drop the short block after a
blank separator when present, otherwise drop the trailing single-line prompt.
Command output is preserved; the fresh shell reprints the live prompt on boot.
An idle terminal tab (no command ever typed) grew one extra copy of the
shell's boot prompt on every close/reopen: persistSnapshot re-serialized a
buffer that was just the replayed old prompt plus the fresh shell's new
prompt, and cleanReviveSnapshot's blank-line trim can't strip prompts on
shells like default PowerShell that print no separator line.
Track real user input (keystrokes/paste, drag-and-drop paths, injected
commands) and, when a session had none, skip re-serializing. If the buffer
we loaded carried no real scrollback (empty or only a repeated prompt),
clear it so the next launch shows a single fresh prompt and any existing
accumulation heals; otherwise leave the prior snapshot untouched so real
history from an earlier active session survives an idle reopen.
Salvages #61584 (activity tracking) and #61577 (clearing content-free idle
buffers) into one path: it also heals already-polluted buffers, counts
drag-and-drop and injected input as activity, and never discards genuine
short command history (only empty/all-identical buffers are cleared).
Co-authored-by: alelpoan <alelpoan@proton.me>
Co-authored-by: liuhao1024 <sunsky.lau@gmail.com>
GatewayRunner._prepare_inbound_message_text's "@" context-reference
block read self._model / self._base_url to resolve the model for
get_model_context_length_async. GatewayRunner never sets either
attribute (copy-pasted from HermesCLI in da44c196b, which does carry
self.model/self.base_url). Every message containing "@" raised
AttributeError inside the try block, silently swallowed by the
surrounding except Exception at debug level, so
preprocess_context_references_async never ran and @file:/@folder:/@diff/
etc. references passed through to the model unexpanded.
Fix: resolve model/provider/base_url via
self._resolve_session_agent_runtime(source=, session_key=,
user_config=), the same session-aware resolution the hygiene
compression block already uses a few hundred lines later in this file.
Also raise the swallow log from debug to warning (with exc_info at
debug) so a future regression here is visible instead of silent.
Third correction, and the load-bearing one. The previous commit put the
"did compaction clear the threshold?" verdict inside should_compress(). But
conversation_loop calls should_compress() TWICE per turn with two different
measures (turn_context.py / conversation_loop.py:1033 and :4789):
* pre-API : request_pressure_tokens -- a rough estimate that can dip BELOW
the threshold
* post-API: real prompt tokens -- which stay above it
So the rough reading reset the strike every turn and the loop never stopped.
Reproduced: 8 compactions in 8 turns under the real two-call pattern, even
with the previous fix applied. (My earlier repro only called should_compress()
once per turn, which is why it looked contained.)
Move the verdict to update_from_response(), the one place that sees the
provider's real prompt_tokens for the just-compacted conversation, guarded by
the existing awaiting_real_usage_after_compression flag so it fires exactly
once per compaction. Real-vs-real: it cannot be fooled by a rough sub-threshold
reading, and (from the previous commit's lesson) never subtracts an estimate
from a real count. should_compress() goes back to a plain threshold test plus
the pre-existing cooldown and anti-thrash guards.
New test test_rough_preflight_reading_does_not_reopen_the_loop drives the real
two-call-per-turn pattern and fails on the prior should_compress()-based commit.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Follow-up to the previous commit, whose futility check was unsound:
incompressible_floor = max(0, display_tokens - pre_estimate)
`display_tokens` is the provider's real prompt count; `pre_estimate` is
`estimate_messages_tokens_rough(messages)`. Subtracting an estimate from a real
count folds the tokenizer skew into "floor" and misreads it as incompressible
overhead. With a 1.6x skew on a 200K window (threshold 150K, true floor 30K):
rough_msgs=253,804 real_prompt=436,086
computed floor = 182,282 <-- mostly skew; exceeds the threshold
after compaction: 401 -> 77 msgs, real prompt = 106,361 (CLEARS 150,000)
verdict: ineffective_count = 1 <-- false positive
Two such passes would permanently disable compaction on a healthy session --
worse than the loop this PR set out to fix.
Move the check into should_compress(), where both sides of the comparison are
the caller's own token count:
* prompt under the threshold -> not thrashing; reset the counter
* a compaction just ran and we are STILL over -> one strike
Real-vs-real, so tokenizer skew can never be mistaken for a floor, and nothing
subtracts an estimate from a real count. compress() now only ever increments the
counter; the reset lives with the one measure the trigger uses.
Adds `test_no_false_positive_under_tokenizer_skew` (the case above) and
`test_counter_resets_once_the_prompt_fits_again` (one failed pass must not
disable compaction forever). Against upstream, 5 of the 7 cases fail; the 2 that
pass are the regression guards, which is the intended shape.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
`should_compress()` documents anti-thrashing protection ("if the last two
compressions each saved less than 10%, skip compression to avoid infinite
loops"). In practice `_ineffective_compression_count` reset on every pass,
so the guard was dead code and a mis-sized context window presented as a
hung CLI instead of a warning.
Two defects:
1. Mixed measurement bases. Effectiveness was
`(current_tokens - estimate(compressed)) / current_tokens`, where
`current_tokens` is the provider's FULL prompt (system prompt + tool
schemas + messages) but `estimate(compressed)` covers messages only.
Every compaction therefore reported ~96% savings and reset the counter.
Savings is now scored messages-vs-messages.
2. Message shrinkage is the wrong yardstick. `should_compress()` trips on
the full prompt, but compaction can only shrink messages -- the system
prompt and tool schemas are an incompressible floor. When that floor
alone meets the threshold, each pass shrinks messages by a healthy
margin, legitimately resets the counter, and still leaves the prompt
over the line; the next turn compacts again, forever. Observed in the
wild: 45+ consecutive compactions, one auxiliary-LLM call each, zero
progress. Effectiveness is now scored against the goal -- did the
projected prompt get under the threshold? -- and a futile pass warns
with the numbers that prove it.
Also record an ineffective pass on the "only N messages (need > M)" early
return, which previously returned the transcript unchanged without moving
any anti-thrash state -- the same class of bug the neighbouring
"no compressable window" branch was already fixed for.
Tests: 4 of the 5 new cases fail on main and pass here; the fifth pins
that effective compaction still resets the counter (121 -> 15 messages,
88.9% savings).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Fold in the TUI direction from #61192 and cover the remaining new-live-session picker path with one shared session-argument normalizer.
Co-authored-by: DatTheMaster <hermesagent424@gmail.com>
Desktop active-session picker calls already pass a session_id, but the gateway's model switch persistence is controlled by parsed model flags. Add --session so the shared parser keeps live-session selections, including MoA virtual provider presets, out of profile config.yaml.
Constraint: config.set model values are parsed by hermes_cli.model_switch before persistence is decided.
Rejected: backend special-case for desktop session_id | it would duplicate existing --session semantics and widen the gateway surface.
Confidence: high
Scope-risk: narrow
Directive: Keep desktop model picker active-session switches explicit with --session; do not rely on session_id alone for persistence.
Tested: npm run test:ui -- src/app/session/hooks/use-model-controls.test.tsx src/app/shell/model-menu-panel.test.tsx
Tested: npm run typecheck
Tested: git diff --check
Not-tested: full pytest suite; change is desktop TypeScript/UI routing only.
Preserve deletability, route identity, stored costs, aggregate reconciliation,
and zero-usage Codex route accounting on top of the salvaged per-model usage
work.
The `sessions` table records only the initial (model, billing_provider)
for a session, so when a user switches models mid-session (via `/model`
or programmatically) every token — including the switched model's — is
attributed to the first model. Insights/billing reports then hide the
cost of the new model entirely (e.g. a session that started on deepseek
and switched to opus shows $0 for opus).
Add a `session_model_usage` table keyed (session_id, model,
billing_provider) that accumulates each per-API-call delta under the
model active at the time of the call. `update_token_counts()` is the
single chokepoint every per-call delta flows through (CLI, gateway,
cron, delegated, codex), so recording there captures accurate
attribution on every platform. Only the incremental path records — the
gateway's `absolute=True` summary overwrite is skipped to avoid
double-counting cumulative totals that can't be split per model. When a
call omits the model, it falls back to the session's recorded model,
matching the existing COALESCE-from-session summary behaviour.
Insights `_compute_model_breakdown` now aggregates tokens and cost from
`session_model_usage`, so a switched session splits correctly across
models, with a defensive fallback to the per-session aggregate for any
session lacking usage rows. A v17 migration backfills one usage row per
existing token-bearing session from its aggregate totals (idempotent via
INSERT OR IGNORE), validated lossless against a 1.3 GB production DB.
Tests: per-model recording, mid-session split, model fallback, absolute
no-double-count, v17 backfill, and an insights-level switch breakdown.
Fixes#51607.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Bundle Fireworks AI as a first-class BYOK provider across the CLI, web/TUI,
and desktop onboarding.
- New model-provider plugin with attribution headers (HTTP-Referer / X-Title)
so Fireworks can attribute Hermes traffic; PAYG-safe default aux + fallback
models (accounts/fireworks/models/...), IDs tracking fw-ai/fireconnect.
- Registered in CANONICAL_PROVIDERS so it appears in the CLI/web/TUI pickers.
- Alias wiring (fireworks-ai, fw) into both CLI resolvers.
- First-class wiring: OPTIONAL_ENV_VARS, HERMES_OVERLAYS (FIREWORKS_BASE_URL
override), doctor env hints. Live catalog + model_metadata are auto-derived.
- doctor: treat Fireworks' native slash-form IDs (accounts/fireworks/...) as
valid, not aggregator vendor prefixes, so it no longer tells Fireworks users
to switch to openrouter or drop the prefix.
- picker: plugin providers with no static curated list now lead with their
profile fallback_models, so the default is an agentic chat model instead of
whatever the live catalog returns first (Fireworks listed an image model,
flux-*, ahead of its chat models).
- Desktop onboarding: Fireworks as a RECOMMENDED hero card with the official
Fireworks logomark and a brand-purple badge, routing to the BYOK key form;
i18n in en/ja/zh/zh-hant.
- Tests: profile contract, first-class wiring (both resolvers, overlay, config,
doctor incl. the slash-form regression, aux headers, credentials), discovery
spot-check, and a live smoke test driven through the Hermes runtime.
Fire Pass (fpk_) support is coming soon; the future wiring is kept as a
commented-out scaffold in the plugin.
`hermes -t web chat` silently dropped the toolset filter (and the same
hold true for `-m`, `--provider`, `--tui`, `--dev` placed before
`chat`). Reported in #28780 for `-t/--toolsets`; the others are sibling
failures with the same root cause.
Root cause: the chat subparser re-declared these flags with `default=None`
(or `default=False` for store_true) on top of the matching top-level
parser flags. When argparse dispatches into the subparser it shares the
namespace via `dest`, so the subparser's default overwrites whatever the
top-level parser parsed before the subcommand. `-s/--skills`, `-r/-c/-w`,
`--yolo`, and `--pass-session-id` already use `default=argparse.SUPPRESS`
for exactly this reason — the chat-subparser action becomes a no-op
unless the user explicitly passes the flag after `chat`, and the parent
value survives.
Reproduction (origin/main, before fix):
>>> parser.parse_known_args(["-t", "web", "chat"]).toolsets
None
>>> parser.parse_known_args(["chat", "-t", "web"]).toolsets
'web'
After fix:
>>> parser.parse_known_args(["-t", "web", "chat"]).toolsets
'web'
>>> parser.parse_known_args(["chat", "-t", "web"]).toolsets
'web'
Sibling flags fixed in the same commit because they share the exact same
argparse pattern bug — verified via a new contract test that scans every
chat-subparser action whose `dest` is also on the top-level parser and
asserts `default is argparse.SUPPRESS`. The test fails on origin/main
listing all five offenders and passes after this fix.
Test additions in tests/hermes_cli/test_argparse_flag_propagation.py:
- TestChatSubparserInheritedValueFlags exercising real `_parser` build
(not the hand-rolled replica) so it catches future drift.
- Parametrized before-chat / after-chat cases for `-t`, `--toolsets`,
`-m`, `--model`, `--provider`.
- Negative case: passing none of the flags leaves attrs at the top-level
parser's `None` default (SUPPRESS does not remove existing attrs).
- Combined case: all three value flags before `chat` simultaneously.
- store_true cases for `--tui` / `--dev`.
- Contract test asserting every shared-`dest` flag on chat uses SUPPRESS.
Fixes#28780.
Review feedback: a same-host redirect to a different port can land on a
different service, which must not inherit the provider API key. Compare
(scheme, hostname, effective port) — with 80/443 defaults — instead of
hostname alone, and add a two-server regression test for the
same-host/different-port case.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
fetch_models() sends Authorization: Bearer <api_key> plus any
default_headers (x-api-key etc.) via urllib.request.urlopen, and
urllib's redirect handler forwards every header when following a
3xx — including to a different host. A catalog endpoint (or a
compromised/misconfigured proxy in front of it) answering with a
redirect to another origin therefore received the provider API key.
Install an HTTPRedirectHandler that drops authorization, x-api-key,
api-key, x-goog-api-key and cookie when the redirect target hostname
differs from the original request, mirroring the pattern already used
in skills/creative/comfyui/scripts/_common.py. Same-host redirects
keep credentials so legitimate path-level redirects still work.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
_CodexCompletionsAdapter (agent/auxiliary_client.py) is a second,
independent producer of Codex Responses input — used by auxiliary
calls (context compression, flush_memories, MoA aggregation,
session_search) that route through CodexAuxiliaryClient instead of
the main agent's ResponsesApiTransport.build_kwargs. It calls
_chat_messages_to_responses_input() directly without is_github_responses,
so the previous commit's fix didn't cover it: an auxiliary call made
against a Copilot-backed session could still replay a connection-scoped
codex_message_items id and hit the same HTTP 401.
Detect the Copilot host from the adapter's own client.base_url (same
check the adapter already does further down for prompt_cache_key
opt-out) and pass is_github_responses through, closing the gap.
Still #32716.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Copilot (api.githubcopilot.com/responses) binds replayed assistant
codex_message_items ids to a specific backend "connection". Credential-
pool rotation, a gateway restart, or routine load-balancer churn between
turns all invalidate that binding, and Copilot rejects the stale id with
HTTP 401 "input item ID does not belong to this connection" — even for
short ids well under the #27038 64-char length cap, since this is a
connection-scope problem, not a length problem. Once a session captures
one of these ids it is persisted and replayed forever, permanently
bricking the session.
Thread an is_github_responses flag from build_kwargs/convert_messages
into _chat_messages_to_responses_input and drop the id unconditionally
on that path, mirroring how reasoning items already strip id on replay.
phase/status/content are still replayed so cache-relevant signal isn't
lost — only the connection-scoped id is unsafe to reuse.
Written to apply independently of the #27038 length-cap fix so the two
PRs don't block each other; they touch adjacent conditions in the same
block and merge cleanly in either order.
Fixes#32716
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Keep empty-tail recovery scoped to the current stream segment and bound fallback flood retries. Preserve Telegram's server retry hint without blocking final delivery through a long cooldown.
Cron jobs in the gateway process wedged before HTTP on later non-streaming
API calls because interruptible_api_call spawned a daemon worker inside
nested cron thread pools. Route cron platform turns through direct_api_call
on the conversation thread instead.
Snapshot the selected stored session and route token for the full async submit
pipeline so a mid-flight session switch cannot resume the wrong chat or
misroute the user's text. Includes regression tests.
Fixes#54527 — a message typed into one TUI session could be silently
misrouted into (or overwritten by) another concurrently-open session.
Root cause: activeQueueSessionKeyRef is written on every render, but the
debounced draft-persist timer, the pagehide flush, and dispatchSubmit's
reject-restore path all read it lazily at async-resolve time instead of
capturing the scope that was active when the operation started. A session
switch landing between capture and resolve relabels one session's text
under the other session's key. A large paste widens the window (slower
synchronous render), which matches the original report.
Fix: introduce draftScopeRef, written only by the draft-swap effect (so it
always reflects the session whose text is actually loaded in the editor)
and read it instead of the render-time ref at both async write sites.
dispatchSubmit's restore() now uses the submittedScope already captured at
dispatch instead of re-reading the live ref.
Also adds isPendingDraftPersistCurrent as defense-in-depth: before the
debounce timer commits a write, it verifies its captured {scope, text}
pair is still the one on file. This is a no-op under the fix above (a
session swap or a newer keystroke already clears/replaces the pending
entry via clearTimeout), but turns any future regression that reintroduces
a stale/live-ref read at this call site into a dropped write instead of a
silent cross-session misroute.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Codex assigns assistant message items server-side ids that can run
400+ chars (base64 encrypted blobs), but the Responses API caps
input[].id at 64 chars and rejects the whole request with a
non-retryable HTTP 400. Once a session captures one of these long
ids, every subsequent turn replays it and 400s forever, since the
history persists it in codex_message_items.
Add a 64-char length guard at both replay sites — the history-to-
input converter and the final preflight gate — so oversized ids are
dropped while short ids (msg_...) are kept for prefix-cache hits.
Mirrors the existing pattern for reasoning items, which already
strip their id before replay because store=False means the API
can't resolve ids server-side anyway.
Fixes#27038
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
scripts/desktop-sandbox.sh runs a Hermes desktop instance in an isolated
sandbox — separate HERMES_HOME, separate Electron userData, and a
distinct
app name (HERMES_DESKTOP_APP_NAME) so it doesn't compete with the main
desktop instance's single-instance lock.
Two modes:
- Ephemeral (default): temp dir, cleaned up on exit
- --persistent: stored under .hermes-sandbox/ in the worktree git root,
survives restarts for repeat testing
In the Nix devShell the script is available as 'sandbox'.
Also makes APP_NAME overridable via HERMES_DESKTOP_APP_NAME in main.ts —
app.setName() runs before requestSingleInstanceLock(), so the overridden
name changes the lock key. collectRelaunchEnv already preserves
HERMES_DESKTOP_* vars through self-update relaunches; test updated to
cover the new env var.
we built .ts into .js for a minute there and dumped em in src, and those
old .js files are getting resolved over top of the ts updated ones so
desktop clients don't update.
just --clean the old files so there's never a conflict :3
The SLASH_TRIGGER_RE regex used (?:^|[\s]) as its left anchor, so typing
a / anywhere in the message (e.g. "hello /") opened the slash command
popover — even though slash commands only execute at the beginning of a
message. Anchor the regex strictly at position 0 (^) so the popover only
appears when / is the first character, matching the actual execution
semantics. The @-mention trigger is left untouched since those work
anywhere in the text.
removing tsc -b from the build script (previous commit) also removed
the only step that type-checked the electron/ directory — the CI
typecheck job runs tsc -p . --noEmit, which uses tsconfig.json whose
include is only ["src", "../shared/src"], so electron/ was silently
uncovered. extend the typecheck script to also run against
tsconfig.electron.json so electron/ stays type-checked in CI.
finalize passed conversation_history=history aliasing the snapshot so flush
skipped every message and wrote nothing. now flush _session_messages via
marker dedup like gateway shutdown. add real db e2e tests.
force_close_tcp_sockets stayed shutdown-only after #29507 to avoid
cross-thread FD recycle. That left CLOSED sockets unreclaimed when
httpx.close() skipped already-shutdown sockets under long-lived
gateways (~1 CLOSED fd / 6 min via proxy).
Add release_fds= for the owning-thread dispose path only; abort still
defaults to shutdown-only.
get_due_jobs()'s one-shot stale-entry recovery (#38758) treated an
expired run_claim (#59229) as proof the claiming tick died, but a run
stalled on network I/O — or a laptop asleep mid-run — legitimately
outlives the TTL while very much alive. The recovery then deleted the
job record mid-flight: list showed the job gone, and when the run
finished mark_job_run() found nothing to update, so last_run_at /
last_status / last_delivery_error were never recorded.
Two guards, per the liveness signals available:
- Same process (the common single-gateway case): before removing a
dispatch-limit-reached one-shot, consult the scheduler's running set
via a lazy import; if the job is still running here it is slow, not
stale — keep the entry.
- Cross process: run_job's monitor loop now refreshes run_claim.at
every 60s while the run is alive (including under
HERMES_CRON_TIMEOUT=0, which previously blocked without polling), so
an expired claim really does mean the owner died and the TTL stays a
dead-owner detector.
Fixes#62002
When web_search results are passed directly to web_extract, the URLs
field contains dict objects (e.g., {"url": "...", "title": "..."})
rather than plain URL strings. Two code paths assumed URLs were always
strings and crashed:
- agent/display.py get_cute_tool_message for web_extract: tried to call
url.replace() on a dict, causing AttributeError
- tools/web_tools.py web_extract_tool loop: tried regex search on a dict,
causing TypeError
Both now extract the URL string from dict objects (url or href field) or
fall back to empty string, preserving the cosmetic display and allowing
the tool to process the URLs correctly.
Fixes#61693
Curator review forks now pass credential_pool and request_overrides from resolve_runtime_provider into AIAgent so pool-backed custom providers can rotate credentials on 401 like main chat.
Add a glyph-agnostic ParticleField (float-up + organic sway/bank + springy
pop-in), skinned as pink pixel hearts. Hearts play on the pet when one is out
(in-window or popped out) and celebrate alongside; otherwise they rise from the
composer. A generic $petReaction bus mirrors the burst to the pop-out overlay
window so it reacts even while the app is minimized.
Consume the core `reaction` event to fire hearts on affectionate messages. DEV
Shift+H previews a burst.
Replace the client-side GOOD_VIBES_RE detection with the backend `reaction`
event: on it, flash the status-bar heart and the pet's celebrate pose. Detection
now lives once in the core, so the TUI, CLI, and desktop stay in sync.
tui_gateway forwards reaction_callback as a `reaction` event (shared by the TUI
and the desktop app). The interactive CLI wires reaction_callback to flash the
pet's celebrate ("jump") pose — the CLI's analogue of hearts.
Add a token-free, curated affection matcher (agent/reactions.py) — the single
source of truth for detecting user "vibes" (ily / <3 / good bot / heart emoji).
No model call, no tokens. Generalized to return a reaction *kind* so future
reactions can ride the same signal.
Wire an opt-in AIAgent.reaction_callback that fires from build_turn_context on
the incoming user message. It never touches the conversation (cache-safe) and
never fatal — a purely cosmetic side-beat each host can consume.
Radix's hoverable-content grace area can leave tips stuck over Electron drag regions; disable it and make tip content pointer-events-none so open state tracks the trigger only.
Switching connection mode (local / cloud agent / remote) no longer
full-window-reloads into the cold-boot CONNECTING screen. The primary
backend is torn down in place (no renderer reload); the shell + Settings
stay up while session lists are wiped so sidebar skeletons retrigger, then
the socket re-dials and config/sessions refresh. Cold-boot CONNECTING
latches off after the first successful boot; the intentional teardown
suppresses the backend-exit toast. Dev affordance: a "Preview soft switch"
button under Gateway diagnostics (Electron has no ?query= entry).
Gateway settings UI brought in line with the rest of Settings:
- Mode cards use the shared selectableCardClass on an equal-height
auto-rows-fr grid, stacking 1→3 (never an orphaned 2+1); titles wrap
instead of truncating.
- Remote gateway's auth detail moves into a ? tooltip in the title; drop
the redundant "connects to the one you choose" from the cloud card.
- textStrong buttons force px-0 so the underline sits flush with the label.
- Tooltip chip uses box-decoration-break: clone so the background hugs each
wrapped line (bg only on the text), capped at max-w-64.
Fully i18n'd (en + zh; ja/zh-hant inherit via defineLocale).
Old packaged Desktop apps re-entering bootstrap against an existing
~/.hermes/hermes-agent were still passing the baked-in --commit pin, which
detached the managed checkout back to the app stamp (e.g. 0.15.1) after
hermes update had already moved it forward.
Skip the packaged commit pin when activeRoot already has git metadata;
keep branch args and fresh-install commit pinning unchanged. Port of
#59902 onto the post-ts-ify bootstrap-runner.ts.
Co-authored-by: helix4u <4317663+helix4u@users.noreply.github.com>
Dashboard Chat is an xterm mirror of a TUI inside the gateway, so
server-side clipboard.paste never sees the browser clipboard. Upload
pasted/dropped images to the profile's images/ dir (same place
clipboard.paste / image.attach use), then drive /image over the PTY.
Uses a dedicated /api/chat/image-upload endpoint (magic-byte check,
25MB cap, profile scope) instead of relative managed-files uploads that
400 on local dashboards without a locked root. Ctrl/Cmd+Shift+V also
tries clipboard.read() for images before falling back to text, since
preventDefault on that chord suppresses the DOM paste event.
Salvages #57912 (client composition + /image PTY drive) and folds in
#48563's upload endpoint + drop path.
Co-authored-by: bird <6666242+bird@users.noreply.github.com>
Co-authored-by: tt-a1i <53142663+tt-a1i@users.noreply.github.com>
Windows drops webContents zoom on minimize/restore, so the UI snapped back
to 100% while Settings still read 125%. Zoom was only reasserted on the main
window's did-finish-load, never on show/restore and never for session windows.
Reassert the persisted level on show/restore + first load, wired once in
wireCommonWindowHandlers so the main window and secondary session windows
share it. The pet overlay opts out (zoom:false): it sizes its own OS window
to fit the sprite in unzoomed CSS px and has its own Alt+wheel scale, so
inheriting the global zoom would render the mascot larger than its window and
crop it (and it shares the renderer origin's zoom localStorage key).
Salvages #61245; keeps its pure-helper tests and adds a scope assertion.
Co-authored-by: HexLab98 <liruixinch@outlook.com>
The second lock block in get_or_create_session held self._lock during six
blocking operations on every inbound message: _is_session_ended_in_db
(SQLite SELECT), _should_reset (callback), _save (SQLite write + JSON write
+ os.fsync), and _recover_session_from_db (SQLite SELECT + UPDATE).
A code comment at line 1607 claimed 'SQLite calls are made outside the
lock' -- true only for _compression_tip_for_session_id, which was moved
out in a prior fix. The remaining I/O was never addressed.
Restructure into a four-phase lock/no-lock split that mirrors the pattern
already established at the bottom of the function:
Phase 1 (lock) -- read entry + session_id
Phase 1b (no lock) -- stale check + reset policy
Phase 2 (lock) -- apply decisions to _entries, capture snapshot + flags
Phase 3 (no lock) -- recovery DB query, _save from snapshot, end/create
_save_entries(snapshot) replaces _save() to avoid dict-mutation races when
called outside the lock. _query_recoverable_session splits the DB I/O out
of _recover_session_from_db so only the _entries assignment needs the lock.
Three early returns inside the lock block are eliminated in favour of a
unified save + return path.
The sync _session_has_compression_in_flight sat on the message hot path
and blocked the event loop twice: under session_store._lock during
_ensure_loaded_locked (JSON read) and via db.get_compression_lock_holder
(SQLite SELECT). Async-ify the method and offload both sources via
asyncio.to_thread; await the call site in _handle_active_session_busy_message.
Every inbound message calls get_or_create_session which synchronously
executes _is_session_ended_in_db → db.get_session → conn.execute on
the asyncio event loop. On a ~1.4GB state.db, this blocks the loop
for seconds to minutes, starving Discord heartbeats.
Upstream #55159 fixed the same pattern for self._session_db in
gateway/run.py but missed SessionStore._db in gateway/session.py.
This follows the exact same approach as #55159:
- session.py internals stay fully synchronous (zero changes)
- Threading.Lock contract is preserved
- All hot-path callers in run.py and slash_commands.py wrap calls
with await asyncio.to_thread(self.session_store.method, ...)
Affected: get_or_create_session, switch_session, update_session,
load_transcript, rewrite_transcript, rewind_session, reset_session,
set_model_override, _save — ~24 call sites across 2 files.
# Conflicts:
# gateway/slash_commands.py
Match the sibling pattern (pet-generate/generate-unavailable.tsx): inline the
portal URL literal in the ExternalLink href instead of a one-off named const.
Drop the portalBaseUrl→IPC→useState plumbing I added for the "create an agent"
link. HERMES_PORTAL_BASE_URL is a dev/staging-only override; threading it
through cloud.status() into React state just to build one link isn't worth it —
in prod it's always portal.nousresearch.com. Module-level constant instead.
Per review: the empty-state "create an agent" link went to the generic portal
agents list; point it at the Hermes Cloud create-instance flow
({portal}/cloud?setup=instance) instead. Derive the host from the portalBaseUrl
that cloud.status() already echoes so it honors HERMES_PORTAL_BASE_URL rather
than hardcoding a second copy of the portal host. Link text/copy → "Hermes
Cloud" (en + zh).
Git Bash hands file tools paths like /c/Users/... which Path() on native
Windows treats as relative \\c\\Users\\... under the process cwd. Reuse
local._msys_to_windows_path (extended for /cygdrive and /mnt drive forms)
in _resolve_path_for_task / _resolve_base_dir so read/write/search land on
the real drive. Container/WSL Linux paths are left untouched.
Salvages #50488 (drops unrelated desktop artifact commit); tests adapted
from #46995.
Co-authored-by: Jeff Watts <186512915+lEWFkRAD@users.noreply.github.com>
Co-authored-by: LeonSGP43 <cine.dreamer.one@gmail.com>
Tighten the salvaged Hermes Cloud code with no behavior change:
- main: one `trimCloudOrg` projection reused by the success-echo and the 409
org list (drop the duplicated map), and a `cloudLoginError()` factory for the
three needsCloudLogin throw sites.
- renderer: a `cloudLoginLapsed()` predicate for the duplicated
needsCloudLogin→signed-out check.
Cleanups on top of @ben's Hermes Cloud salvage:
- isConnectedAgent normalized both sides of the cloud-URL comparison (trim +
drop trailing slash + lowercase). The saved URL is host-lowercased by
normalizeRemoteBaseUrl but the discovered dashboardUrl is raw from NAS, so
a host-casing difference could silently break the connected-highlight.
- DesktopCloudStatus.signedIn doc said "AT-or-RT"; it actually reflects the
Nous portal Privy session (privy-token), not the gateway cookies.
Adds a third "Hermes Cloud" gateway mode to the desktop app: one portal
sign-in auto-discovers the agents on your account and connects to any of
them with no second interactive prompt.
- Electron: widen connection mode to 'local' | 'remote' | 'cloud', routed
through a centralized modeIsRemoteLike() so every resolution site treats
cloud exactly like remote; portal discovery (GET /api/agents over the
OAuth partition), Privy-cookie liveness, multi-org picker (NAS 409), and a
silent per-agent /oauth cascade (load protected root, not /login).
- Persist a cloudOrg on the cloud block; unselect cloud on mode switch.
- Renderer: Hermes Cloud ModeCard + agent picker (signed-out/loading/empty/
list), org picker, Change-org, connected-highlight + Connected pill.
- i18n (en + zh full; ja/zh-hant inherit via defineLocale), Cloud icon.
- IPC: hermes☁️{status,login,logout,discover,agent-sign-in}.
Salvage of #55402 onto current main: the original branch predates the
desktop electron .cjs -> .ts migration (39d09453f), so the electron half
was re-authored against the .ts files. Authorship preserved.
cloud-auto-discovery Phases 3 + 4.
Only restore held verification text when the loop genuinely ends through budget exhaustion. Preserve later interrupts and failures, keep generated-summary fragment explanations, and add regression coverage for both contracts.
Treat intent-ack continuation text as non-final so last-turn exhaustion requests a real summary instead of surfacing a premature promise. Keep iteration-limit fallback text free of the abnormal-fragment explainer.
Name the continuation fallback for its actual verification-only provenance so unrelated continuation paths cannot accidentally inherit its cron-delivery semantics.
Track held-back verification responses explicitly so budget exhaustion returns the composed report without a second model call. Keep unrelated error and recovery exit reasons intact, preserve Kanban timeout accounting, and cover the real run_conversation paths.
Clear stale final_response before verify-on-stop/pre_verify loop
continues, and normalize iteration-limit exit reasons in turn_finalizer
when a composed answer survives with unknown/budget_exhausted.
Fixes#61631
Sibling sites of the salvaged #55997 fix, all reading user-editable
config values through .get(key, '').method(): MoA slot provider/model
labels, gateway quick-command alias targets (2 sites), gateway.proxy_url,
and gateway.relay_url. Regression tests for the contributor's two sites
plus the MoA labels.
dict.get(key, default) returns None (not the default) when the key
EXISTS with value None. The default only applies when the key is ABSENT.
Chained method calls (.strip(), .upper(), .count()) crash with
AttributeError on NoneType.
Fix two confirmed hits:
- auxiliary_client.py: custom provider base_url/api_key (config null)
- anthropic_adapter.py: text block content (API null response)
Pattern: .get(key, "").method() → (.get(key) or "").method()
Without this UA tag the Feishu server does not push group @mention events
over the WebSocket transport. The "channel" tag tells the server to use
the Channel protocol which enables group-message routing in addition to P2P
direct messages.
Root cause: FeishuWSClient was created without any UA signaling tag, so the
server defaulted to the basic DM-only push mode. Group @mention events were
silently dropped before reaching Hermes.
Fixes https://github.com/NousResearch/hermes-agent/issues/50656
Also adds a regression test verifying the UA tag is present in the
FeishuWSClient constructor call.
Sibling of the salvaged #53196 read-path fix: setdefault() does not
replace a present-but-null key, so saving platform tools with
known_plugin_toolsets: null in config.yaml crashed on indexing None.
When config.yaml has known_plugin_toolsets set to null (or any value
mapped to None by the YAML loader), config.get returns None (dict.get
only falls back to the default when the key is absent, not when its
value is None). The subsequent set(known_map.get(platform, [])) then
crashes with TypeError: NoneType object is not iterable and the gateway
fails to start, even though no plugin toolsets are configured.
Add or-empty-dict and or-empty-list guards so a null/None value is
treated as empty instead of crashing the platform-tools resolver.
`_load_web_config()` is typed `-> dict` but returned `load_config().get("web",
{})`, which is `None` when the config has a present-but-null `web:` section
(YAML `web:` with no body). Every caller then does
`_load_web_config().get(...)` and raises `AttributeError: 'NoneType' object
has no attribute 'get'` — this hits `_get_backend`, `check_web_api_key`, and
the extract-char-limit reader.
Separately, `check_web_api_key()` read the backend as
`.get("backend", "").lower()`; a null `web.backend` value yields `None` (the
`""` default only applies when the key is absent), so `None.lower()` raised.
`check_web_api_key` is the `check_fn` gate for `web_search`/`web_extract`, so
this surfaced as an exception during tool-availability checking.
- Make `_load_web_config()` honor its `-> dict` contract (`... or {}`), fixing
the null-`web:`-section crash at every call site.
- Guard the backend value in `check_web_api_key` with `or ""`, mirroring the
existing guard in `_get_backend`.
Adds regression tests for both the null-backend-value and null-web-section
cases.
* test: deflake CI and dev-machine flaky tests in bulk
Fixes ten distinct flake sources found by mining recent CI failures and
running the full suite on a dev machine with real user state:
CI-observed races:
- tests/conftest.py live-system guard: allow signal 0 (pure liveness
probe) through _guarded_kill/_guarded_killpg. psutil.pid_exists()
probes a just-killed grandchild reparented to init; the subtree check
fails for it and the guard RuntimeError'd
test_entire_tree_is_sigkilled_not_just_parent intermittently on
unrelated PRs.
Hermeticity flakes (fail on dev machines with real state, pass on CI):
- agent/coding_context.py: _marker_root() now skips the shared temp
root (tempfile.gettempdir()) like it skips $HOME — a stray
/tmp/package.json flipped every tmp_path test into the coding
posture (9 failures in test_coding_context.py).
- test_agent_guardrails.py: pin MAX_CONCURRENT_CHILDREN=3 via autouse
monkeypatch instead of freezing the user's real config value at
import time (import-time vs call-time config mismatch).
- test_web_tools_config.py: TestCheckWebApiKey now neutralizes the
ddgs package probe and registry providers — the optional ddgs
package in a dev venv lit up the fallback backend.
- test_credential_pool.py: block claude_code/hermes-oauth credential
autodiscovery in the two pool-merge tests that assert exact id
lists (a real ~/.claude/.credentials.json seeded an extra entry).
- test_modal_sandbox_fixes.py: clear _permanent_approved /
_session_approved — the user's real command_allowlist silently
approved the guard-escalation commands under test.
- test_setup_irc.py: stub prompt_checklist to select only the IRC row;
the non-TTY cancel fallback re-ran the real configured platforms'
interactive setup_fn, which hit input() under captured stdin.
- test_doctor.py: TestGitHubTokenCheck now patches the module-level
HERMES_HOME constant (the file's established pattern) instead of
only setenv — doctor was running PRAGMA integrity_check against the
real multi-GB state.db and blowing the 300s per-file budget.
Latent atexit-duplication (same _enter_buffered_busy class as #34217):
- test_undo_command.py: drop importlib.reload(tui_gateway.server) in
fixture teardown; reload re-registers the module's atexit hooks.
- test_session_platform_resolution.py: drop per-test reload of
tui_gateway.server; every resolver reads env at call time.
* test: sentinel model value in ignore-user-config fallback assertion
With HERMES_IGNORE_USER_CONFIG=1, load_cli_config() falls back to the
repo-root cli-config.yaml (untracked, gitignored). On a dev machine that
file can legitimately set the same popular model the test hardcoded
(anthropic/claude-sonnet-4.6), flipping the != assertion locally while
CI (no cli-config.yaml) stayed green. Use an impossible sentinel model
name instead.
Sibling sites of the salvaged #47334 fix: xai/openai/elevenlabs/gemini/
mistral/piper/neutts subsection reads in transcription_tools.py and
tts_tool.py used .get(key, {}) which passes a present-but-null value
through as None. All provider-subsection reads now use .get(key) or {}.
Providers without a DEFAULT_CONFIG entry (e.g. stt.xai) were still
receiving None even after the load_config() deep-merge fix, since the
merge can only fill sections that have defaults.
When stt.local, tts.edge, or other config subsections are explicitly set
to null in config.yaml (which happens by default on a fresh --voice
setup), stt_config.get('local', {}) returns None instead of {} because
YAML null preserves the key. The chained .get('model') then crashes
with 'NoneType' object has no attribute 'get'.
Apply the defensive (x or {}) pattern to every place a config subsection
is read via .get('xxx', {}). Covers local, edge, openai, mistral, and
elevenlabs subsections in both transcription_tools.py and tts_tool.py.
Closes#47318
skills: null crashed with AttributeError, and a bare scalar
disabled: my-skill was split into a set of characters. Both now
normalize the same way agent.skill_utils._normalize_string_set does:
null -> empty set, scalar -> single-item set. Non-dict skills
sections are ignored.
Closes#13026.
_load_context_cache() returned None when context_length_cache.yaml
contained 'context_lengths:' (no value) — YAML parses this as
{'context_lengths': None} and dict.get(key, default) only returns
the default when the key is absent, not when the value is None.
This caused AttributeError in every downstream caller (issue #47135).
Fix: use 'or {}' instead of default= so both absent key and
None value return an empty dict.
Fixes#47135
Addresses Copilot review on #61348: the HTML-escaped role, while safe from
injection (quotes are escaped), still contains whitespace when a crafted role
is supplied, which splits the class attribute into several unintended CSS
classes. Keep the escaped role for the display badge, and reduce the raw role
to a single safe CSS token (alnum/-/_) for the class name. Real roles
(user/assistant/system/tool) are unchanged, so the existing .message-<role>
rules still match.
The gateway reconnect watcher forwards is_reconnect=True to every
adapter.connect() call on every retry. Adapters whose signature omits
the kwarg raise TypeError at every reconnect attempt and stay silently
disconnected — the exact bug that shipped for QQAdapter and only
surfaced after messages stopped flowing on the QQ channel for hours.
This test statically parses every adapter.py under gateway/platforms/
and plugins/platforms/ (via AST, so third-party SDKs like slack_sdk,
matrix-nio, aiohttp, telegram, etc. are NOT required in the test env)
and asserts every *Adapter class with an async connect() accepts
is_reconnect — either as a keyword-only argument or absorbed by
**kwargs.
Also fixes plugins/platforms/wecom/callback_adapter.py:WecomCallbackAdapter,
which the new test caught as a second offender. Same class of bug: bare
'async def connect(self)' signature would die on the first reconnect.
Companion to #59429 (which fixed the original QQAdapter offender).
The base adapter's signature was updated to include
, which the reconnect watcher passes as
during reconnection. All other platform adapters were
updated, but QQAdapter was missed, causing:
TypeError: QQAdapter.connect() got an unexpected keyword argument 'is_reconnect'
This leads to an infinite retry loop since every reconnect attempt fails
immediately with the same TypeError.
Fix: add to QQAdapter.connect()'s signature.
QQBot has no server-side update queue, so the flag is accepted only for
interface conformance.
Test: new test_connect_accepts_is_reconnect_param verifies both
adapter.connect() and adapter.connect(is_reconnect=True) succeed without
raising.
--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 "$PR_NUM" --auto --squash || true
- name:Wait for merge, auto-close on failure or stale
env:
GH_TOKEN:${{ secrets.AUTOFIX_BOT_PAT }}
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."
# A previous push changed the lockfile but the latest one
# doesn't — update the comment rather than leave stale info.
printf '%s\n✅ package-lock.json changes from an earlier push have been reverted — locked versions now match the target branch.\n' "$MARKER" > /tmp/lockfile-diff.md
else
echo "No lockfile changes and no existing comment — nothing to do."
exit 0
fi
fi
if [ -n "$EXISTING" ]; then
echo "Updating existing comment ${EXISTING}"
gh api --method PATCH "repos/${REPO}/issues/comments/${EXISTING}" \
-F body=@/tmp/lockfile-diff.md > /dev/null \
|| echo "::warning::Could not update PR comment (expected for fork PRs — GITHUB_TOKEN is read-only)"
else
echo "Creating new comment"
gh api "repos/${REPO}/issues/${PR}/comments" \
-F body=@/tmp/lockfile-diff.md > /dev/null \
|| echo "::warning::Could not post PR comment (expected for fork PRs — GITHUB_TOKEN is read-only)"
@@ -491,18 +491,18 @@ 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.cjs` + `backendSupportsServe()` in `main.cjs`) 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. Route desktop bugs to the `hermes-desktop-app-work` skill, not `hermes-dashboard-work`.
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`.
**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_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.
- **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.
-`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.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/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.
**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).
**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).
---
@@ -1094,14 +1094,16 @@ kanban task.
- **CLI:** `hermes_cli/kanban.py` wires `hermes kanban` with verbs
@@ -67,6 +67,8 @@ npm run dev # Vite renderer + Electron, which boots the Python backend
Point the app at a specific source checkout, or sandbox it away from your real config:
```bash
# throwaway HERMES_HOME, separate Electron userData, distinct app name to avoid the single-instance lock
../scripts/dev-sandbox.sh npm run dev
HERMES_DESKTOP_HERMES_ROOT=/path/to/clone npm run dev
HERMES_HOME=/tmp/throwaway npm run dev
npm run dev:fake-boot # exercise the startup overlay with deterministic delays
@@ -85,7 +87,63 @@ Installers are built and uploaded to GitHub Releases manually. macOS/Windows sig
### How it works
The packaged app ships the Electron shell and a native React chat surface. On first launch it can install the Hermes Agent runtime into `HERMES_HOME` (`~/.hermes`, or `%LOCALAPPDATA%\hermes` on Windows) — the **same layout a CLI install uses**, so the two are interchangeable. Backend resolution first honours `HERMES_DESKTOP_HERMES_ROOT`, then a completed managed install, then a probed `hermes` on `PATH` (unless `HERMES_DESKTOP_IGNORE_EXISTING=1` is set), and finally an explicit `HERMES_DESKTOP_HERMES` command override for packagers/troubleshooting. The renderer (React, in `src/`) talks to a headless backend the app launches for you — a `hermes serve` process that serves the `tui_gateway` JSON-RPC/WebSocket API — through the framework-agnostic client in [`apps/shared`](../shared/) (the same client the web dashboard consumes), and reuses the agent runtime rather than embedding `hermes --tui`. The app is **self-contained**: it runs its own `hermes serve` backend and never opens or requires the web dashboard UI. (For backward compatibility, a runtime that predates the `serve` command automatically falls back to a headless `dashboard --no-open` — see `electron/backend-command.ts` — so mid-upgrade installs never break.) The install, backend-resolution, and self-update logic all live in `electron/main.ts`.
The packaged app ships the Electron shell and a native React chat surface. On
first launch it can install the Hermes Agent runtime into `HERMES_HOME`
(`~/.hermes`, or `%LOCALAPPDATA%\hermes` on Windows), using the same layout as a
CLI install.
The app has three boundaries:
- **Electron** resolves and validates a runnable backend, owns native
filesystem/git/window capabilities, and exposes a narrow preload bridge.
- **React** owns the Desktop routes, panes, interaction state, and
`@assistant-ui/react` transcript.
- **Hermes Agent** runs as a headless `hermes serve` process and exposes the
`tui_gateway` JSON-RPC/WebSocket API. The renderer connects through
[`apps/shared`](../shared/), which is also used by the browser dashboard.
Backend resolution is an ordered ladder:
1. `HERMES_DESKTOP_HERMES_ROOT`
2. the current source checkout during development
3. a completed managed install
4. `HERMES_DESKTOP_HERMES`, or `hermes` on `PATH`
5. a system Python that can import the Hermes runtime
6. the first-launch bootstrap installer
Candidates are probed before use; an existing shim or interpreter is not enough.
A runtime that predates `serve` falls back to headless
`dashboard --no-open`. This is compatibility for the backend command only and
does not launch or embed the dashboard UI.
The Electron orchestration entry point is `electron/main.ts`; pure resolution,
probe, hardening, and platform policies live in focused modules beside it. The
renderer is under `src/`, with shared atoms in `src/store` and transport/native
adapters in `src/lib`.
Before changing the app, read:
- [`AGENTS.md`](./AGENTS.md): architecture, state ownership, resolver/fallback,
transport, performance, and testing rules.
- [`DESIGN.md`](./DESIGN.md): visual system, information architecture, motion,
direct manipulation, and keyboard behavior.
### Connections, projects, and switching
Desktop supports a managed local backend, explicit remote gateways, and Hermes
Cloud connections. Remote and cloud modes use the same remote-capability path;
authentication and discovery differ, not the renderer feature model.
Projects are the workspace abstraction. A project may own multiple folders,
repositories, worktrees, and sessions; a bare new chat remains detached unless
the user enters a project or configures a default project directory. Use the
Projects UI rather than adding a second per-session folder-picker workflow.
Changing profiles or connection modes is a soft workspace switch, not another
cold boot. The shell and current management overlay remain mounted while
gateway-bound nanostores are wiped, query-backed data is invalidated, and the
new connection repopulates skeletons. This prevents rows or transcripts from
the previous gateway bleeding into the next one.
### Verification
@@ -95,9 +153,13 @@ Run before opening a PR (lint may surface pre-existing warnings but must exit cl
npm run fix
npm run typecheck
npm run lint
npm run test:desktop:all
npm run test:ui
npm run test:desktop:platforms
```
Run `npm run test:desktop:all` for install, boot, update, packaging, or other
release-path changes.
### Troubleshooting
Boot logs land in `HERMES_HOME/logs/desktop.log` (includes backend output and recent Python tracebacks) — check it first if the app reports a boot failure.
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
ethernet