Repository navigation
feat(review): a Reviewer agent reviews every change before it is tested - #264
Merged
Merged
Conversation
Multi-agent for real on the Coding → Testing hand-off: after a proposed change's checks go green, the Reviewer agent (read-only, on the member's computer, scoped autonomy) reads the diff in context against the ticket and the repo's own rules — correctness, security / tenant isolation, tests, conventions — and grades findings must-fix / should-fix / nit. It posts the review on the pull request; must-fix findings become the change request the Coding agent gets next (the same 3 bounded rounds as failing checks and failed acceptance), then the change is reviewed again. An approved change goes on to the Testing agent and Autopilot's merge policy. - Domain: IReviewAgent, ReviewVerdict / ReviewFinding (an "approve" with a must-fix finding is not trusted). - Pipeline: ReviewAgent (prompt, JSON verdict, PR comment, change request). - Web: BoardTaskService.ReviewAsync; repo states Reviewing / Reviewed / ChangesRequested (session counts them as Done); AutopilotPlanner RunReview + ChangeReview; orchestrator maps states, recovers a review cut short by a restart; plain-language statuses on the Board. - Docs: autonomy-levels phase table. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018WAdQ7Rm24xss5QDSvXtJB
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Answering "is this really multi-agent?": the hand-off between the Coding and Testing agents now has a Reviewer agent.
Flow per proposed change: checks green → Reviewer agent (read-only, scoped, on the member's computer) reads the diff in context against the ticket + repo rules (correctness, security / tenant isolation, tests, conventions), grades findings must-fix / should-fix / nit, and posts the review on the PR → approved: Testing agent → merge policy; must-fix: back to the Coding agent as a change request (same 3 bounded rounds), then reviewed again.
IReviewAgent,ReviewVerdict/ReviewFinding— an "approve" carrying a must-fix is not trusted.ReviewAgent(prompt, verdict parse, PR comment, change request), DI.BoardTaskService.ReviewAsync; repo statesReviewing/Reviewed/ChangesRequested; session recompute treats them as Done.AutopilotPlanner:ChangeReview+RunReview; orchestrator maps states, recovers a review cut short by a restart.Test plan
ReviewStepTests(13): review before testing, needs a computer, waits while reviewing, changes go back with findings, 3-round cap, approved → test → merge; verdict parsing incl. untrusted approve; change request; comment escaping.🤖 Generated with Claude Code
https://claude.ai/code/session_018WAdQ7Rm24xss5QDSvXtJB