Gemini 3.1 Pro Preview may emit parts with only thoughtSignature
and no text content, causing the stream to stall. Emit a
thinking_signature event to keep the stream active, and start
a thinking block when these parts arrive before any text.
Fixes#76071
Detect run-level timeouts that fire while a tool call is still active and keep them out of assistant model fallback.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Summary:
- The PR updates the Discord plugin with REST priority lanes and stale background drops, gateway send/member v ... normalization, explicit component forwarding, one-off component wait helpers, tests, and a changelog entry.
- Reproducibility: yes. The prior scheduler failure has a high-confidence source-level reproduction using one ... ne pending same-bucket request under fake timers, and the latest head adds a regression test for that path.
ClawSweeper fixups:
- Included follow-up commit: fix(discord): preserve option localizations
- Included follow-up commit: fix(discord): harden component sends
- Included follow-up commit: docs(changelog): note Discord Carbon parity hardening
- Included follow-up commit: fix(discord): harden Carbon parity
- Ran the ClawSweeper repair loop before final review.
Validation:
- ClawSweeper review passed for head 4fa634ef1a.
- Required merge gates passed before the squash merge.
Prepared head SHA: 4fa634ef1a
Review: https://github.com/openclaw/openclaw/pull/75363#issuecomment-4357438917
Co-authored-by: Peter Steinberger <steipete@gmail.com>
Co-authored-by: clawsweeper <274271284+clawsweeper[bot]@users.noreply.github.com>
* fix: raise Gemini 2.5 minimal thinking budget from 128 to 512
Google raised the minimum thinkingBudget for Gemini 2.5 Pro and Flash
models to 512. The hardcoded minimal: 128 now causes a 400 INVALID_ARGUMENT
("The thinking budget 128 is invalid. Please choose a value between 512
and 24576.") on every heartbeat using gemini-2.5-flash-lite with minimal
thinking.
* test(google): assert thinkingBudget minimum of 512 for Gemini 2.5 models
Add parametrized tests covering minimal/low/medium budget tiers for
gemini-2.5-flash-lite, gemini-2.5-flash, and gemini-2.5-pro — ensures
the 128→512 floor change is regression-tested.
Pre-existing typecheck failure in src/agents/model-auth.ts (TS2322) is
unrelated; confirmed present on upstream main without our changes.
Tests validated: pnpm test extensions/google/transport-stream.test.ts (21 pass).
* test(google): fix it.each format string placeholder order
* fix(google): narrow Gemini 2.5 minimal budget floor to flash-lite only
Add Control UI timing attribution for tab visibility, refresh phases, secondary refresh groups, and gateway RPCs. Decouple slow Overview, Cron, and Nodes secondary refreshes from primary tab feedback, and report Cron runs timing from the controller's real ok/error/skipped load outcome.
Also keeps telemetry secret-safe by excluding request params and response bodies, adds focused regression coverage for RPC attribution and stale/failed background refreshes, and carries the latest-main Discord test fixture correction needed for check:test-types.
Fixes#64004.
Summary:
- This PR adds Discord outbound mention-format guidance to the Discord plugin prompt hints, Discord channel docs, a focused channel test, and the changelog.
- Reproducibility: not applicable. as a runtime bug reproduction. There is a high-confidence inspection path: ... lacks the guidance, while the existing formatter/tests and Discord reference establish the expected syntax.
ClawSweeper fixups:
- Included follow-up commit: fix(discord): document mention formatting guidance
Validation:
- ClawSweeper review passed for head bf2734a002.
- Required merge gates passed before the squash merge.
Prepared head SHA: bf2734a002
Review: https://github.com/openclaw/openclaw/pull/75173#issuecomment-4354537199
Co-authored-by: Peter Steinberger <steipete@gmail.com>
Co-authored-by: clawsweeper-repair <clawsweeper-repair@users.noreply.github.com>
Co-authored-by: Peter Steinberger <58493+steipete@users.noreply.github.com>
