* feat(clawhub): surface rate-limit headers and sign-in hint on 429s
* feat(clawhub): handle no-headers 429s and add changelog entry
* fix(clawhub): simplify 429 message to reset hint only
When a newer inbound message arrives while an older dispatch is still
processing, buffered final answers from the old dispatch could be
delivered into the new message's thread.
Add `bufferedGeneration` to `BufferedFinalAnswer` to track which
abort fence generation buffered the answer. `takeBufferedFinalAnswer`
now rejects buffered answers whose generation doesn't match the
current dispatch's generation, preventing stale content delivery.
Fixes#76642
Keep Mantis Discord comparison worktrees outside the uploaded artifact tree and copy each lane's QA output into the report directory before comparing summaries.
Defers automatic PDF model/auth resolution until the PDF tool is executed, avoiding agent-turn tool prep stalls while preserving explicit PDF model registration.
Fixes#76644.
Thanks @hclsys.
Document that channels.telegram.streaming.preview.toolProgress requires
channels.telegram.replyToMode: 'off'. Quote-reply requires the final message
reference at send time, which is incompatible with preview-edit streaming, so
the two features are mutually exclusive on Telegram.
Adds:
- Note callout in docs/channels/telegram.md after the existing toolProgress
guidance, explaining the exclusion and how to restore visibility.
- Cross-link bullet in docs/concepts/streaming.md pointing to the Telegram
channel doc for the full note.
Surfaces a doc/runtime gap that has been silent since v2026.4.22.
* fix(discord): normalize whitespace in command description comparison
Discord server-side storage collapses consecutive whitespace and removes
whitespace between adjacent CJK characters when persisting slash-command
descriptions. Our locally-serialized desired descriptors keep the original
whitespace, so commandsEqual returned false on every startup for any
deployment with multi-line or CJK-heavy descriptions.
This caused reconcile to issue a PATCH for every such command on every
gateway restart. Under Discord's per-application rate limit that quickly
produced a burst of 429s and some commands silently failed to register
until the next restart — a recurring symptom behind several Discord deploy
429 reports (#75341, #75888).
Fix: apply the same two normalizations to description strings in
comparableCommand — collapse runs of whitespace to a single space, then
drop whitespace between CJK boundary characters — before JSON-based
equality. This mirrors Discord's storage semantics so round-tripped
descriptions compare equal.
Adds command-deploy.test.ts with regression coverage for:
- server-side default fields ignored (dm_permission, nsfw, version, ids)
- required:false treated as absent (existing behavior)
- CJK descriptions with \n separators normalized
- mixed CJK/ASCII descriptions with consecutive whitespace
- substantive description diffs still flagged
- ASCII whitespace normalization
* fix(discord): normalize localized command descriptions
* fix(discord): ignore null localization maps in command reconcile
---------
Co-authored-by: Peter Steinberger <steipete@gmail.com>
Summary:
- Adds an OpenCode provider-policy public artifact that delegates Claude thinking profiles, plus regression tests and a changelog entry for preserving `xhigh` on `opencode/claude-opus-4-7`.
- Reproducibility: yes. Source inspection on current main shows the model-switch path remaps unsupported store ... vider-policy-api` artifact for the fallback resolver; the proposed test exercises that stored-`xhigh` path.
Automerge notes:
- No ClawSweeper repair was needed after automerge opt-in.
Validation:
- ClawSweeper review passed for head efa152cca5.
- Required merge gates passed before the squash merge.
Prepared head SHA: efa152cca5
Review: https://github.com/openclaw/openclaw/pull/76760#issuecomment-4366483080
Co-authored-by: mkdev11 <MkDev11@users.noreply.github.com>
Summary:
- The branch changes Telegram polling startup to reuse the successful probe `getMe` result as grammY `botInfo` ... es` after recoverable `deleteWebhook` failures, and updates Telegram docs, changelog, and regression tests.
- Reproducibility: yes. for the narrow PR bug: source inspection shows current main can block before polling o ... d timeout coverage that reaches `run()`. The full linked high-RTT report remains only partially reproduced.
Automerge notes:
- Ran the ClawSweeper repair loop before final review.
- Included post-review commit in the final squash: fix(telegram): start polling after webhook cleanup timeout
- Included post-review commit in the final squash: fix(telegram): extract bot info contract
Validation:
- ClawSweeper review passed for head c74bbdd1ff.
- Required merge gates passed before the squash merge.
Prepared head SHA: c74bbdd1ff
Review: https://github.com/openclaw/openclaw/pull/76735#issuecomment-4366417178
Co-authored-by: Ayaan Zaidi <hi@obviy.us>
Co-authored-by: Peter Steinberger <steipete@gmail.com>
Summary:
- Adds a Vitest regression case in `src/agents/session-file-repair.test.ts` for `repairSessionFileIfNeeded` preserving a final stop assistant message after a tool-call/tool-result pair.
- Reproducibility: yes. The added JSONL sequence can be traced through current `repairSessionFileIfNeeded`: no ... ry, so the function returns `repaired: false`; I did not execute the test because this review is read-only.
Automerge notes:
- No ClawSweeper repair was needed after automerge opt-in.
Validation:
- ClawSweeper review passed for head ee3aa0e036.
- Required merge gates passed before the squash merge.
Prepared head SHA: ee3aa0e036
Review: https://github.com/openclaw/openclaw/pull/76538#issuecomment-4365670008
Co-authored-by: Vinh Nguyen <vinhnt36@fpt.com>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Fixes#76176.
OpenAI live verification showed `gpt-5.4-mini` supports reasoning effort generally, but rejects `/v1/chat/completions` payloads that combine function tools with `reasoning_effort`. This keeps reasoning effort for tool-free Chat Completions and Responses, and omits it only for the rejected Chat Completions + function tools combination.
Validation:
- Live OpenAI API matrix on 2026-05-03
- pnpm test src/agents/openai-reasoning-effort.test.ts src/agents/openai-transport-stream.test.ts -- --reporter=verbose
- GitHub PR CI green on ea3915308c
Thanks @ThisIsAdilah and @chinar-amrutkar.
Fixes the session-list plugin metadata hot path by reusing the active workspace-scoped plugin metadata snapshot for manifest model-id normalization and setup CLI backend fallback. Also keeps model-pricing collection on its provided manifest registry and fixes the CI-only hoisted test mock failure.
Validated with targeted plugin/gateway/model-selection tests, CI-shaped gateway startup shard, Testbox `pnpm check:changed`, and green PR CI.
Thanks @rolandrscheel.
