fix(ci): filter ClawSweeper comment dispatches before token minting#95308
Conversation
|
Codex review: needs maintainer review before merge. Reviewed June 21, 2026, 7:01 AM ET / 11:01 UTC. Summary PR surface: Tests +25, Config +18, Other +10. Total +53 across 6 files. Reproducibility: yes. from source inspection: current main creates the target comment token for every issue_comment before the later grep exits on non-command comments. I did not run a live GitHub Actions reproduction in this read-only review. Review metrics: 1 noteworthy metric.
Stored data model Merge readiness Overall follows the weaker of proof and patch quality, so missing proof can cap an otherwise strong patch. Rank-up moves:
Risk before merge
Maintainer options:
Next step before merge
Security Review detailsBest possible solution: Keep the narrow prefilter, preserve generic github_activity forwarding for ordinary comments, and merge only after maintainers verify both non-command and supported-command issue_comment behavior on the exact head. Do we have a high-confidence way to reproduce the issue? Yes from source inspection: current main creates the target comment token for every issue_comment before the later grep exits on non-command comments. I did not run a live GitHub Actions reproduction in this read-only review. Is this the best way to solve the issue? Yes, this is the narrow target-repo fix for the OpenClaw ingress side: move command detection before target write-token minting while leaving generic activity forwarding intact. The remaining best-fix requirement is live dispatch proof against the receiver contract. AGENTS.md: found and applied where relevant. Codex review notes: model internal, reasoning high; reviewed against b796890b9777. Label changesLabel changes:
Label justifications:
Evidence reviewedPR surface: Tests +25, Config +18, Other +10. Total +53 across 6 files. View PR surface stats
What I checked:
Likely related people:
What the crustacean ranks mean
Shiny media proof means a screenshot, video, or linked artifact directly shows the changed behavior. Runtime, network, CSP, and security claims still need visible diagnostics. How this review workflow works
|
Dependency graph guard clearedThis PR no longer has blocked dependency graph changes. A future dependency graph change requires a fresh
|
|
Merged via squash.
Thanks @vincentkoc! |
Summary
Pre-filter issue comments before minting the target ClawSweeper App token or dispatching comment work. Ordinary comments still reach the generic activity path, but only comments containing a ClawSweeper mention or supported slash command mint the write token and run the command acknowledgement/dispatch path.
Verification
This is intentionally scoped to the OpenClaw ingress workflow; the corresponding ClawSweeper-side rate-limit hardening is in the companion PR.