sweep: accept a spoken noop token when codex skips the file write#44
Merged
Merged
Conversation
codex occasionally ends with STUPIFY_NO_NEW_ISSUES as its final message without writing the review file (seen on bevyl #7528/#7537/#7627); the run then read as FAILED and the head was throttled for an hour. Parse the final codex message (last 'codex' line to its 'tokens used' footer, never the diff-bearing transcript) and treat an exact token match as the noop it meant to write.
Octember
force-pushed
the
claude/stupify-status-re-reviewing-220c09
branch
from
July 18, 2026 02:41
c0a8747 to
5154ff8
Compare
![exe-dev-github-integration[bot]](https://avatars.githubusercontent.com/in/3061187?s=60&v=4){kind=small}
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
codex sometimes finishes a clean review by saying
STUPIFY_NO_NEW_ISSUESas its final message instead of writing it to the review file (observed on bevyl #7528, #7537, #7627 in the 2026-07-17 sweep log). The runner only reads the file, so those runs loggedreview FAILEDand the head got throttled forFAIL_RETRY_MIN(60 min) — PRs looked ignored.Fix: when the review file is empty but the run succeeded, parse the final codex message (the text between the last bare
codexline and itstokens usedfooter) and accept an exactSTUPIFY_NO_NEW_ISSUESmatch as the noop it meant to write. Only that message can stand in for the file — never the whole transcript, which inlines the untrusted diff — and a run with notokens usedfooter (truncated) is still a failure.Tested with the real transcript shape from the VM's sweep.log, plus injection and truncation cases.