Commit Graph
20 Commits
Author SHA1 Message Date
Rohit C Prasad 5f2eeca1c8 Diagnose truncated tool calls instead of executing their mangled args
Unparseable (_raw) args now get a truthful error: cut-off-by-output-limit says
'smaller pieces', bad JSON says 're-send with declared parameters'. Raw junk is
shrunk before entering history so replays can't teach the model the _raw shape.
Anthropic default max_tokens 16k -> 32k so typical report files fit outright.
2026-08-15 06:55:10 -07:00
Rohit C Prasad 560fc3cb8a Tool-request card speaks plainly; declining re-checks for a user-installed copy
Fact strip: 'OpenWorker installs its own verified copy from <publisher>' replaces
supply-chain jargon. On decline the engine re-resolves — a copy the user installed
themselves is handed to the agent as theirs, not treated as a refusal.
2026-08-15 00:14:15 -07:00
Rohit C Prasad 8e77d61aa1 Tool-request card: separate the product's facts from the coworker's ask
Registry metadata (version, publisher, checksum) moves to a distinct fact strip.
Decline button renamed to say the run continues; reason capped to one sentence.
2026-08-14 23:50:51 -07:00
Rohit C Prasad 25d32891d3 Tool-request prompts fail closed on installability
TOOL_REQUESTED now carries the registry's verdict (installable/version/summary).
GUI offers Install only when the event says a pinned build exists.
2026-08-14 23:25:21 -07:00
Rohit C Prasad 62ad9dbdca tools: give coworkers the user's real toolchain, and stop silent skips
Sidecar inherits the login shell's env; toolchain resolves absolute paths with pinned
installs; request_tool replaces the 'tool missing -> STOP' instruction that hid a check.
2026-08-14 15:28:57 -07:00
Rohit C Prasad 06a4365132 approvals: session read-only command grant; enable on consent card
'Allow read-only commands' on shell cards — fail-closed classifier (local reads + pipelines only; no network/interpreters/writes), session-scoped, persisted with grants.
Consent cards get an in-place Enable button.
2026-08-11 15:07:46 -07:00
Rohit P 53d5fc498f Merge branch 'main' into issue/ope-51-ask_user-upgrades 2026-08-01 09:45:03 -07:00
Devika Verma 70e4610a3e Add support for Skills (#391)
Global & per-workspace skill.
Per-persona skills will be improved later as we re-design that abstraction, as per roadmap.
2026-08-01 08:50:53 -07:00
Rohit C Prasad f9f51c97c6 compaction: live progress signal + user-message cap
COMPACTING event drives a 'Compacting context…' transient in the GUI.
Cap the compacted block's user-message list at 40 with an honest omitted count.
2026-07-30 06:24:39 -07:00
Devika Verma f08a3c425b compaction: engine hook, failure policy, persistence (OPE-27 2/4)
Minimal engine footprint: a checkpoint at each iteration top (between tool
turns and before a new turn), the usage signal captured per round-trip
(context_tokens; chars/4 estimate when never reported), and
_outbound_messages consulting the boundary. The summarizer runs off-loop
through the normal provider router, so the Settings model pin is just an
id.

Failure policy per spec: retry once in both modes; attended sessions get
the Retry / Trim-oldest-10% prompt (via the ask_user plumbing, gated by an
is_attended callback the WS surface wires); unattended runs auto-trim and
continue — never parked on internal bookkeeping. Raw context-overflow 400s
from the main model route into the same policy, progress-guarded so a
still-overflowing model terminates in the error path.

CompactionState persists on the session record (new sqlite column, same
defensive parse as grants), so reloads keep the compacted view. A
persisted compacted notice + a new COMPACTED event mark the spot for
the GUI divider (rendered in commit 3).
2026-07-29 16:13:14 +05:30
Devika Verma 70cd1fa3d4 ask_user upgrades: rich options, grouped questions, option previews (OPE-51)
Options accept {label, description, recommended, preview} objects (plain
strings unchanged — old sessions render as today's pills), and `questions`
groups up to 4 questions into one call, rendered as a stepper via the
header chips. Any option preview switches the card to a two-pane layout:
options left, monospace pane right, following hover/focus.

Grouped calls resolve with a JSON map keyed by header-or-question and
return {answers: {...}} to the agent (single stays {answer: ...});
a grouped item's first question doubles as its title/options so channel
mirrors and legacy surfaces degrade sensibly. Channel buttons use option
labels; grouped items mirror as text with the open-the-app hint.
2026-07-29 15:52:05 +05:30
Rohit C Prasad 979badbd3c Meter token usage across all model providers
Normalized TokenUsage (input/output/cache split) captured in every provider's
stream and complete paths, persisted as an assistant-message sidecar and sent
on the assistant_message event; matrix gains verified context-window sizes.
2026-07-27 21:01:31 -07:00
Rohit C Prasad f2b8fde428 Keep Retry offered across model switches
Switch notices no longer consume the retry guard or hide the button.
Error, switch model, Retry is the intended recovery path; retry runs on the new model.
2026-07-23 07:43:24 -07:00
Rohit C Prasad 0e8b85e6f3 Show reasoning traces: live Thinking block + persisted disclosure
New reasoning_delta event; traces persist as a display sidecar stripped from provider feeds.
Sources: compat vendors' reasoning_content and Gemini thought summaries (include_thoughts).
Live-verified on GLM via Together and Gemini 3; Gemini tool loops stay healthy.
2026-07-22 16:15:11 -07:00
Rohit C Prasad f1eb652d61 Allow mid-session model switching with a persisted transcript marker
Picker stays live for the session; switches persist a model_switch notice (§17 revised).
Rebinds refused mid-turn; images become placeholders for non-vision targets at send time.
2026-07-22 15:43:54 -07:00
Rohit C Prasad a63710ecfe Add Gemini 3 models with thought-signature support
Signatures ride the assistant message as a _gemini sidecar and are reattached in tool loops.
Thought-flagged parts are filtered from answer text; foreign sidecars stripped on the OpenAI wire.
Live-verified both Gemini 3 models end-to-end; without the echo they 400 on every tool loop.
2026-07-22 14:05:23 -07:00
Rohit C Prasad 878b858ece Persist error/interrupt markers in history; add Retry on failed turns
Engine appends a display-only notice message on error/interrupted; providers never see it.
New retry frame re-runs a failed turn with no new user message, guarded on the error tail.
GUI renders persisted notices on reload and a Retry button on the trailing error.
2026-07-22 13:45:43 -07:00
Rohit C Prasad d8903bc927 Keep interrupted partial streams in the transcript
GUI flushes the streaming buffer into a durable item on interrupted/error.
Engine persists partial text on the provider-error path like the stop path.
e2e red-green verified; full suites pass.
2026-07-22 10:40:01 -07:00
Rohit C Prasad f6ad97274d engine: Stop works in every state, not just between iterations
Stream drops between chunks (+ the pre-first-token wait); pending approvals/questions/plans resolve as interrupted.
Running shell commands die via an executor interrupt hook; skipped tool calls still get results (no orphans).
Also fixes delete_session calling a nonexistent engine.interrupt().
2026-07-21 23:16:40 -07:00
Rohit C PrasadandDevika 2b45018ffa OpenWorker: initial import
Imported from andrewyng/aisuite@1b4bbf303e
(contents of its platform/ directory, hoisted to the repo root).
Development history prior to this commit lives in that repository.

Co-authored-by: Devika <devikaverma11@gmail.com>
2026-07-21 11:09:41 -07:00