632 Commits

Author	SHA1	Message	Date
Vincent Koc	d60584ee2d	test(google): cover ready callback close precedence	2026-08-02 01:25:41 +08:00
Vincent Koc	814a2ebb61	fix(google): flush lazy prompts on provider readiness	2026-08-02 01:22:33 +08:00
Vincent Koc	3a8afb2a71	Merge origin/main into fix/google-realtime-lazy-queue-bounds-20260801 ... * origin/main: (44 commits) fix(plugins): compare authoring metadata by canonical JSON semantics (#117486) fix(doctor): reject contradictory lint check selectors (#117495) fix(anthropic): honor configured model pricing (#117275) refactor(plugins): reuse scoped loader for capability discovery (#117373) refactor(test): derive changed tooling routes (#117407) fix(kova): clarify extension memory attribution (#117500) fix(imessage): sanitize private context on every outbound path (#117501) fix(ci): align release validation fixtures (#117494) fix(talk): preserve owner authorization in realtime consults (#114163) fix(ui): surface aborted-run diagnostics (#117497) perf(models): skip Ollama row runtime loading (#117465) fix(cli): report user profile command outcomes (#117466) fix(active-memory): detect Chinese recall intent in escalate mode (#117338) (#117419) fix(workboard): ignore caller-supplied archivedAt when creating a card (#116412) fix(models): avoid waiting on promotions before listing models (#117322) refactor(plugins): inherit shared package boundary settings (#117474) fix(systemd): parse all inline environment assignments (#117484) feat(web-fetch): add tools.web.fetch.headers for operator request headers (#115545) fix(qa): Slack progress lane misses successful delivery (#117391) refactor: deduplicate Codex prompt snapshot fixtures (#117476) ...	2026-08-02 01:02:06 +08:00
Vincent Koc	877843422e	fix(google): avoid dropping first realtime voice input (#117450) ... * commit '2ccb11ac28daea1939e6666c1a0a3d346159d9fe': test(google): satisfy realtime live lint test(google): wait for realtime response drain test(google): diagnose realtime live readiness test(google): prove realtime ready owns live session test(google): cover realtime setup activation ordering fix(google): gate realtime readiness on session ownership	2026-08-02 00:59:02 +08:00
Vincent Koc	9cb647e735	test(google): prove lazy realtime prompt delivery	2026-08-02 00:15:41 +08:00
Peter Steinberger	568ea29d2e	refactor(plugins): inherit shared package boundary settings (#117474)	2026-08-01 09:11:09 -07:00
Jesse Merhi	d3f4530ce8	feat(google): allow operator headers on Gemini web search (#115549) ... * feat(google): support Gemini web search headers * fix(google): validate Gemini search headers * docs(changelog): credit Gemini headers * test(secrets): route config contracts through collector * chore: remove release-owned changelog entry * test(google): cover resolved search headers	2026-08-02 01:41:09 +10:00
Vincent Koc	85d959ee61	test(google): satisfy realtime live lint	2026-08-01 22:58:29 +08:00
Vincent Koc	b1a0ee8539	test(google): wait for realtime response drain	2026-08-01 22:51:35 +08:00
Vincent Koc	b2320c84ee	fix(google): bound lazy realtime user messages	2026-08-01 22:47:12 +08:00
Vincent Koc	6f3f662a86	test(google): diagnose realtime live readiness	2026-08-01 22:46:14 +08:00
Vincent Koc	11d53bebfe	test(google): prove realtime ready owns live session	2026-08-01 22:36:06 +08:00
Vincent Koc	9ba8ffad5d	test(google): cover realtime setup activation ordering	2026-08-01 22:33:02 +08:00
Vincent Koc	ee61f725f1	fix(google): gate realtime readiness on session ownership	2026-08-01 22:33:02 +08:00
Vincent Koc	fcc20d14a3	fix(google): stop scraping Gemini CLI OAuth credentials (#117167) ... * fix(google): delegate Gemini OAuth refresh to the CLI * refactor(google): remove retired Gemini OAuth stack * chore(ci): prune retired Google OAuth baseline * fix(google): retire Gemini CLI usage telemetry * fix(google): remove retired usage token parser	2026-08-01 11:36:24 +08:00
Vincent Koc	541593dcb4	merge: refresh google live transcript bound onto main ... * commit '7e78de747b8a64d52f30aee948b697bf4c528691': (359 commits) refactor(ui): replace duplicated locale bundles with canonical translation memory (#117157) fix(ui): retire stale Gateway event-log observers (#117169) fix(agents): keep subagent status consistent after restart and large fan-outs (#117141) refactor(ui): unify filtered session list ownership (#117158) refactor(agents): consolidate context budgets and compaction recovery (#117149) fix(net-policy): redact sig and x-* auth params in URLs and bodies (#116957) refactor(agents): remove duplicate generated-media delivery fallbacks (#117150) refactor(reply): unify turn lifecycle state ownership (#117145) fix(ui): preserve emoji agent avatar initials (#104912) refactor(auto-reply): unify command and directive ownership (#117143) refactor(plugins): consolidate descriptor and startup ownership (#117146) fix(messages): reply actions leak citation markers and reply/poll answers draw the no-reply fallback (#116909) fix(slack): let durable ingress retry transient thread lookups (#117135) refactor(doctor): consolidate shipped state migration ownership (#117142) fix(ai): Codex stream shows internal parser text on a malformed frame (#116966) fix: guard every migrated session accessor path (#117140) fix(terminal): measure unicode display width consistently (#117062) fix(openai): clear ChatGPT SSE fallback per session (#117123) fix(mistral): reject incomplete streamed tool terminals (#117137) fix(ai): preserve structured chat content and refusals (#117136) ...	2026-08-01 10:45:53 +08:00
Peter Steinberger	3cadc08593	fix(google): restore canonical Gemini and Vertex provider invariants (#117042) ... Co-authored-by: Peter Steinberger <steipete@macos.shared>	2026-07-31 14:31:34 -07:00
Peter Steinberger	d2047364a8	fix(google): preserve streamed termination and usage facts (#116822) ... Co-authored-by: Peter Steinberger <steipete@macos.shared>	2026-07-31 05:29:40 -07:00
Vincent Koc	c163f5272a	merge: refresh google live transcript bound onto main ... * origin/main: (52 commits) fix(setup): isolate failed channel probes and report doctor diagnostics (#116824) fix(telegram): preserve edit previews, message cache, and group history (#116818) fix(whatsapp): preserve interactive replies and normalize media MIME (#116816) fix(file-transfer): keep fetched media attachable in sandboxed replies (#116400) fix(browser): preserve doctor JSON failure status (#116811) fix(ollama): stream native tool call lifecycle (#116809) fix(matrix): preserve message-tool room thread routing (#116802) fix(ui): restore mobile navigation and accessible usage filters (#116751) fix(ui): coalesce profile identity refresh (#116764) fix(ui): preserve chat attachments and run ownership (#116752) fix(plugins): clean up partial service startup once (#116804) fix(qa-channel): enforce message and thread lifecycle ownership (#116801) fix(qa-lab): support groups and isolate thread timelines (#116803) feat(plugins): externalize Voyage embeddings (#116785) fix(outbound): strip echoed inbound metadata before delivery (#50520) fix(anthropic-vertex): correct multi-region endpoints (#116757) fix(ui): scope agent cron jobs and surface model catalog failures (#116771) fix(ui): preserve scoped session state across updates (#116772) fix(build): fail closed on symlinked build output roots (#116705) fix(commands): validate native settings and preserve Telegram ingress outcomes (#116773) ...	2026-07-31 20:20:01 +08:00
Peter Steinberger	0d40cfbb53	fix(google): surface blocked Gemini prompts (#116756) ... Co-authored-by: Peter Steinberger <steipete@macos.shared>	2026-07-31 04:29:06 -07:00
Vincent Koc	74918607d1	test(google): cover live transcript overflow	2026-07-31 18:20:45 +08:00
Vincent Koc	4f65b7e8fe	fix(google): bound live transcript accumulation	2026-07-31 18:20:45 +08:00
Vincent Koc	0ce230c861	test(google): cover expired CLI OAuth compatibility	2026-07-31 16:40:38 +08:00
Vincent Koc	091bc7f9ed	Merge remote-tracking branch 'origin/main' into experiment/google-cli-supported-paths ... * origin/main: (115 commits) test: type session request parameters (#116713) fix: retain event invalidation for pagination docs: note queued session refresh fix fix: preserve queued session refresh options fix(exec): require approval for abbreviated inline eval flags [AI] (#116529) fix: keep owner-only tools out of non-owner skill commands [AI] (#116532) fix(tui): require a fresh agent roster (#116715) fix(memory-wiki): preserve user edits when rolling back ChatGPT imports (#116517) fix(agents): cron failures preserve timeout and cancellation reasons (#116566) fix(release): unblock beta validation workflows (#116708) fix(ci): discover frozen target startup benchmark cases (#116711) fix(session): preserve reset message payloads (#116195) fix(qa): tolerate loaded-runner bus delivery stalls (#116448) improve(ci): rebalance embedded agent test configs (#116436) fix(browser): return committed trace path (#116702) fix(qa): exceed Slack table aggregate limit in fallback probe (#116447) fix(qa): use built plugin tools server (#116442) fix(test): keep memory FD repro config valid (#116698) fix(telegram): keep queued tool progress in one message (#116685) fix(qa): stop waiting on yielded fanout runs (#116440) ... # Conflicts: # src/system-agent/setup-inference-detect.ts # ui/src/e2e/model-setup.e2e.test.ts	2026-07-31 16:40:33 +08:00
Vincent Koc	326a4831cc	test(google): cover realtime connection ownership	2026-07-31 11:47:59 +08:00
Vincent Koc	cc62f37d6e	fix(google): own realtime connection attempts	2026-07-31 11:47:59 +08:00
Vincent Koc	9e965073e5	Merge remote-tracking branch 'origin/main' into experiment/google-cli-supported-paths ... * origin/main: (168 commits) fix(plugins): keep registry sqlite root consistent (#116427) fix(cli): keep Gateway auth failures parseable in JSON (#116597) fix(trajectory): preserve prompts in oversized events (#116220) test(buzz): add live QA Lab channel lane (#116298) chore(i18n): refresh native locales (#116567) feat: add OpenClaw TUI prototype skill (#116583) fix(update): accept concrete versions selected by ranges (#116585) fix(agents): preserve locked session model routes test(qa-lab): preserve Matrix fault proxy identity fix(ci): extend OpenAI live profile setup budget (#116580) fix: prevent configured models from stalling startup (#116553) fix(cli): close MCP listener after one-shot commands fix(auto-reply): keep explicit NO_REPLY silent (#116548) test(discord): cover deferred backpressure terminals fix(discord): cancel backpressured realtime playback test(discord): cover realtime output backpressure fix(codex): add remote WebSocket heartbeat and reconnect (#116135) fix(macos): wait for the current reconnect snapshot (#116550) fix(outbound): recognize current-source message-tool sends (#116560) fix(voice-call): type cancelled consult outcomes ... # Conflicts: # extensions/google/setup-api.test.ts # ui/src/e2e/model-setup.e2e.test.ts	2026-07-31 09:30:03 +08:00
Josh Avant	ce67ffb70e	feat(agents): add tool-free isolated completion (#114343) ... * feat: add isolated pure-inference completion * fix(google): block ambient system prompt writes * docs: refresh generated map	2026-07-30 13:24:45 -05:00
Vincent Koc	c7d1205d51	fix(google): guide broken CLI profiles to supported auth	2026-07-30 22:40:20 +08:00
Vincent Koc	c568f93b03	fix(google): retire Gemini CLI OAuth setup	2026-07-30 16:33:33 +08:00
Vincent Koc	0370fce7ff	fix(test): isolate Google embedding batch timeout (#116161)	2026-07-30 10:11:29 +08:00
Peter Steinberger	5bfc65d7f4	refactor: remove 3,543 lines of redundant runtime and tests (#115961) ... * refactor: remove 3,543 lines of redundant runtime and tests * refactor: ratchet production environment variable budget	2026-07-29 11:43:41 -04:00
Monkey-wusky	79e0b559fe	fix(google): cancel unread response body on video download error (#110045) ... * fix(google): cancel unread response body on video download error * fix(google): safely cancel failed streaming video downloads --------- Co-authored-by: Peter Steinberger <steipete@gmail.com>	2026-07-29 09:00:04 -04:00
Peter Steinberger	4232126bba	chore: update dependencies across workspace (#115677) ... * chore(deps): update dependencies * fix(deps): restore CI compatibility	2026-07-29 05:16:42 -04:00
Peter Steinberger	ddd0b48673	refactor: remove 3,527 lines of runtime and test duplication (#115626) ... * refactor: remove obsolete runtime and test duplication * test: preserve typed cron and realtime fixtures	2026-07-29 02:21:20 -04:00
Vito Cappello	33dae97b81	fix(agents): restore subagent completion delivery on CLI runtimes (#115422) ... * fix(agents): deliver tool-free CLI completions * fix(agents): enforce empty Gemini CLI tool caps * refactor(agents): remove ineffective Gemini admin settings * fix(agents): block inherited Gemini MCP servers * fix(agents): enforce Gemini MCP cap via argv --------- Co-authored-by: VACInc <3279061+VACInc@users.noreply.github.com> Co-authored-by: Patrick Erichsen <patrick.a.erichsen@gmail.com>	2026-07-28 21:07:02 -05:00
Peter Steinberger	d5907ca619	fix(google): keep static compat flags on discovered Gemini id variants (#114831)	2026-07-27 21:13:44 -04:00
Peter Steinberger	b5bb29a36e	fix(code-mode): preserve direct tools and secure suspended runs (#114778) ... Preserve trusted policy-required message delivery and native image/computer tools across Code Mode catalogs and provider transports. Keep preferred Gemini capability metadata during live discovery, escape MCP declaration keywords, and reject suspended-run resumes with missing owner identity. Add before-and-after regressions for each user path. Closes #114753.	2026-07-27 18:47:33 -04:00
Peter Steinberger	7394262cd1	feat: per-model code-mode capability flags with an "auto" master-switch tier (#114695) ... * feat(tools): per-model code-mode capability flags and auto master-switch tier * feat(anthropic): flag claude-sonnet-4-6 as code-mode preferred * revert(anthropic): drop claude-sonnet-4-6 from the code-mode preferred set * test(anthropic): type the manifest compat field in the catalog contract test * fix(anthropic): carry catalog compat onto hand-built forward-compat model rows	2026-07-27 16:47:41 -04:00
Peter Steinberger	17ed69e103	fix(agents): preserve prompt caches across tool discovery (#114572)	2026-07-27 09:58:19 -04:00
Peter Steinberger	91b1f9676f	fix(media): reject malformed provider binary payloads (#114514)	2026-07-27 06:39:47 -04:00
zengLingbiao	a8e52b49a0	fix(google): reject malformed base64 audio in Gemini TTS responses (#114041) ... * fix(google): reject malformed base64 audio in Gemini TTS responses * fix(media): validate provider audio payloads Co-authored-by: zengLingbiao <zeng.lingbiao@xydigit.com> * fix(xai): terminate malformed realtime audio Co-authored-by: zengLingbiao <zeng.lingbiao@xydigit.com> --------- Co-authored-by: Peter Steinberger <steipete@gmail.com>	2026-07-27 04:17:35 -04:00
Peter Steinberger	8e4b6623e5	test(providers): consolidate streaming and model contract cases (#114336)	2026-07-27 01:36:32 -04:00
Peter Steinberger	caa01f4f5b	refactor(providers): table-driven forward compat and manifest-derived contract tests (#113940) ... * refactor(providers): share family forward compat resolution * test(providers): derive runtime contracts from manifests * test(providers): validate manifest fixture inputs * refactor(zai): narrow forward compat inputs	2026-07-25 18:09:37 -07:00
Masato Hoshino	bdcc5d54a5	fix(agents): honor run abort signal in image and pdf tools (#112644) ... * fix(agents): honor run abort signal in image and pdf tools The image and pdf agent tools declared `execute: async (_toolCallId, args)` and dropped the run abort signal that `wrapToolWithAbortSignal` supplies as the third execute argument. The wrapper only races the execute promise, so an aborted run kept sequentially downloading images/PDFs (up to the per-tool cap, each up to the byte cap) and still issued a paid vision/PDF-model call for a dead run. Thread the signal into the existing `requestInit: { signal }` seam (which the media fetch layer already merges into the download fetch) and add `signal.throwIfAborted()` between sequential loop items and before the paid model call. No new media-options signal field; non-abort behavior is unchanged. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix(agents): keep the pdf test suite under the lint ceiling; guard model dispatch Follow-up on the abort-signal change, resolving the check-lint failure and a review finding on the same seam. max-lines --------- Adding the two pdf abort tests pushed src/agents/tools/pdf-tool.test.ts to 1018 effective lines against a ceiling of 1000, failing check-lint. Resolved without touching config/max-lines-baseline.txt — a suppression there would also have tripped the max-lines ratchet. - The new tests declared their own `describe` with `beforeEach`/`afterEach` hooks identical to the existing `describe("createPdfTool")`. They exercise that same tool, so they now live in it and the duplicated scaffolding is gone. - `stubPdfToolInfra`, `createPdfModelRegistry` and `FAKE_PDF_MEDIA` moved to pdf-tool.test-support.ts, which already exists for exactly this. They go through `createPdfToolInfraStub(completeMock)` rather than being exported directly, because the stub wires the suite's own `complete` mock into the model registry and vi.mock handles are file-scoped. All 21 existing call sites are unchanged. Abort propagation into model dispatch ------------------------------------- The previous revision stopped cancellation at download boundaries and before the first model call, but `runImagePrompt`/`runPdfPrompt` never saw the signal. A run cancelled while the first provider request was in flight could still issue the remaining ones — the image path dispatches `describeImage` once per image in a sequential loop, so a dead run kept paying for every later image. Both now take an optional `signal` and check it immediately before each provider dispatch (3 sites in image-tool, 4 in pdf-tool). Forwarding the signal further, into the provider transports themselves, is a different seam and is deliberately left out of this PR. Also declares `requestInit` on `ImageToolLoadWebMediaOptions`. The local facade omitted it while the underlying loader accepts it (web-media.ts declares it and forwards it to readRemoteMediaBuffer), so the option worked at runtime and only compiled because spread properties skip excess-property checking. A non-spread call site would not have. tsgo core + core-test, oxlint, oxfmt clean; pdf-tool and image-tool suites 266 tests pass. * test(agents): prove in-flight media aborts * test(agents): narrow PDF loader options * test(agents): reject abort mocks with errors * fix(agents): propagate media cancellation * test(agents): type PDF abort fixture * test(agents): type prepared runtime snapshot * fix(agents): normalize abort rejection errors --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Co-authored-by: Peter Steinberger <steipete@gmail.com>	2026-07-25 05:17:09 -07:00
Peter Steinberger	b0603cb443	refactor(fs): adopt shared safe filesystem primitives (#113596) ... * refactor(fs): adopt fs-safe helpers * fix(zalouser): keep stable per-profile QR temp path * fix(fs): route adoption through owned facades	2026-07-25 03:24:19 -07:00
Vincent Koc	749dbcb2a7	refactor(media): consolidate generated media byte limits (#113556)	2026-07-25 15:39:49 +08:00
joshavant	a325467557	fix(cron): preserve scheduled caps across CLI and worker upgrades	2026-07-24 01:24:03 -05:00
joshavant	8287e21445	fix(cron): enforce caps across CLI backends	2026-07-24 01:24:03 -05:00
Peter Steinberger	acd92f6a3d	chore(deps): refresh repository dependencies (#112453) ... * build(deps): update QA broker dependency * build(deps): refresh repository dependencies * build(deps): reconcile rebased shrinkwraps * test(plugins): remove stale loader test state * test(deps): stabilize updated dependency coverage * fix(swift): use caller-isolated TaskLocal overload * build(deps): regenerate rebased shrinkwraps * test(msteams): preserve DNS validation in fetch helper * fix(deps): avoid vulnerable optional image stack * test(deps): validate generated LRU override * refactor(ui): extract chat resizable divider * test(ui): update divider ownership path * fix(matrix): retain restart-compatible SDK * style(cron): format update test	2026-07-23 16:17:13 +00:00