You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Session Handoff [auto-1425]: Detect CodeRabbit's Command-Reply Rate Limit in pr_review.py for #1425 #3127
Fix pr_review.py Detects Only CodeRabbit's Auto-Review Rate Limit, Not the One It Returns to an Explicit Review Command #1425: make scripts/pr_review.py detect CodeRabbit's rate-limit reply to an explicit @coderabbitai review command, and surface it the same way the auto-review notice already surfaces, through rate_limited_by and the digest's other_rate_limited= field. Today RATE_LIMITED matches only the auto-generated comment: rate limited by <service> --> marker. The command-reply shape is an auto-generated reply by CodeRabbit marker followed by an Action not completed block reading Review rate limited., and it names no service in the marker, so attribute the service from the commenting login (a known reviewer in OTHER_REVIEWERS) rather than from the marker. Keep the reading structural (markers and the block's fixed text), never free prose, per the comment above RATE_LIMITED. Done looks like a pull request into develop carrying Closes on promotion: #1425, with tests in tests/test_pr_review.py feeding a constructed command-reply comment that now reads as rate limited, the existing auto-review marker still reading as before, and a CodeRabbit comment that is neither (for example an ordinary review summary) reading as not rate limited. Surface it the same way means the same field and the same reading. The issue's side question, whether a limit with no stated retry time should read differently from one with a countdown, is not part of this lane, so leave that reading unchanged and do not file it as done.
External blockers
None known.
Internal dependencies
None. #2252 also edits pr_review.py's reviewer reading, for Qodo, and needs a design decision first, so it does not belong in this lane.
State
No branch or worktree exists for this lane yet. develop is thirteen squash commits ahead of main, none of them naming #1425. No open pull request exists in the repository.
Parked decision queue
None.
What the last round did
The unattended picker filed this link from the open backlog.
Merged: Read CodeRabbit's Command-Reply Rate Limit in pr_review.py #3128 into develop as squash commit ddd7ddb. scripts/pr_review.py now reads CodeRabbit's command-reply rate limit, the auto-generated reply marker over an Action not completed / Review rate limited. block. It surfaces that through rate_limited_by and the digest's other_rate_limited= field, the same reading the auto-review marker gives. The service is named by login through RATE_LIMIT_SERVICE. Tests in tests/test_pr_review.py cover the reply shape, the existing marker, near misses, and an unmapped login.
Out of scope, as the handoff stated: whether a limit with no retry time should read differently. That reading is unchanged.
Cleanup: the worktree and local branch are removed, the remote branch is deleted by pr_review.py cleanup, and the base clone is fast-forwarded to develop.
Activity
ptr727 commentedon Oct 16, 2026
Lane closed out, its work complete.
scripts/pr_review.pynow reads CodeRabbit's command-reply rate limit, theauto-generated replymarker over anAction not completed/Review rate limited.block. It surfaces that throughrate_limited_byand the digest'sother_rate_limited=field, the same reading the auto-review marker gives. The service is named by login throughRATE_LIMIT_SERVICE. Tests intests/test_pr_review.pycover the reply shape, the existing marker, near misses, and an unmapped login.Closes on promotion: #1425, so it closes when develop is next promoted to main. This run's scope was develop, so no promotion was opened.pr_review.py cleanup, and the base clone is fast-forwarded to develop.