Conversation
Iterative review-and-fix cycle that reviews current diff, triages findings as fix/skip, applies fixes, verifies with lint+test, and repeats until clean (max 3 iterations). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add an iterative review-and-fix command that spawns 5 parallel review agents (correctness, security, performance, conventions, free review), triages findings, fixes actionable items, verifies with lint+test, and repeats until clean (max 3 iterations). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
⚡ Performance Benchmark
Details
|
|
Important Review skippedAuto incremental reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Pro Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
📝 WalkthroughWalkthroughA new markdown file is added that documents an iterative code review workflow process. The workflow defines repeating cycles (maximum 3) of parallel multi-agent reviews across categories (Correctness, Security, Performance, Conventions, Free), followed by triage and selective fix application with verification steps. Changes
Estimated code review effort🎯 1 (Trivial) | ⏱️ ~3 minutes Possibly related PRs
🚥 Pre-merge checks | ✅ 3✅ Passed checks (3 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Summary of ChangesHello, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed! This pull request adds a new command-line tool, Highlights
🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console. Using Gemini Code AssistThe full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips. Invoking Gemini You can request assistance from Gemini at any point by creating a comment using either
Customization To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a Limitations & Feedback Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for GitHub and other Google products, sign up here. Footnotes
|
![devin-ai-integration[bot]](https://avatars.githubusercontent.com/in/811515?s=60&v=4){kind=small}
PR Review: chore(cli): Add review-loop commandOverall: Looks good ✅ — This is a small, self-contained addition of a new agent command file. The prompt is well-structured and the review angles are complementary. FeedbackMinor suggestions (non-blocking)
Premortem Analysis
VerdictClean, well-scoped PR. The command follows the established patterns of other |
Deploying repomix with
|
| Latest commit: |
e2fca26
|
| Status: | ✅ Deploy successful! |
| Preview URL: | https://2d5cb03e.repomix.pages.dev |
| Branch Preview URL: | https://chore-add-review-loop-comman.repomix.pages.dev |
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request introduces a new review-loop custom command by adding a markdown file that defines its behavior. The command orchestrates an iterative review-and-fix cycle using multiple specialized agents, which is a comprehensive approach. The process is well-defined, but I have one suggestion to enhance the clarity of the verification step's instructions to prevent potential ambiguity for the agent executing the command.
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (1)
.agents/commands/code/review-loop.md (1)
16-16: Allow skipped items to be reconsidered if new changes affect them.“Do not re-raise skipped items” is a bit too absolute; skipped findings can become valid after later edits. Consider reopening only when newly changed lines alter the original context.
Suggested wording tweak
-5. **Re-review** only the newly changed lines. Do not re-raise skipped items. +5. **Re-review** only the newly changed lines. Do not re-raise skipped items unless new changes materially affect their context.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.agents/commands/code/review-loop.md at line 16, The sentence "Re-review only the newly changed lines. Do not re-raise skipped items." is too absolute; update the rule to allow previously skipped items to be reopened if later edits affect their context. Replace that line (the "Re-review only..." sentence) with wording that requires reviewers to re-review only changed lines but permits reopening skipped findings when new changes touch or alter the original context (for example: re-review changed lines and reopen skipped items only when new edits affect their context). Ensure the new text appears where the current sentence is located and references the same rule heading ("Re-review") so intent remains clear.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In @.agents/commands/code/review-loop.md:
- Line 11: Agent 4 — Conventions currently points to CLAUDE.md instead of the
project’s canonical rules; update the conventions agent to load and use the
project's canonical rules source (base.md) for checks. Locate the "Agent 4 —
Conventions" entry in the review-loop config and replace the CLAUDE.md reference
with the canonical rules file (base.md), ensuring the agent reads and applies
that ruleset during triage so conventions checks use the project-specific
source.
---
Nitpick comments:
In @.agents/commands/code/review-loop.md:
- Line 16: The sentence "Re-review only the newly changed lines. Do not re-raise
skipped items." is too absolute; update the rule to allow previously skipped
items to be reopened if later edits affect their context. Replace that line (the
"Re-review only..." sentence) with wording that requires reviewers to re-review
only changed lines but permits reopening skipped findings when new changes touch
or alter the original context (for example: re-review changed lines and reopen
skipped items only when new edits affect their context). Ensure the new text
appears where the current sentence is located and references the same rule
heading ("Re-review") so intent remains clear.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
Run ID: ff6f42cb-83fb-4a39-a589-2ebd57555158
📒 Files selected for processing (1)
.agents/commands/code/review-loop.md
- Reference .agents/rules/base.md instead of CLAUDE.md for conventions - Make verify step explicit about repeating until checks pass Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
PR Review Update: chore(cli): Add review-loop commandSince my previous review, the code hasn't changed. Adding incremental feedback on AI bot comments for maintainer triage. AI Bot Inline Comment EvaluationCodeRabbit — Line 11: Conventions agent rule reference
CodeRabbit's actionable comment suggests the conventions agent references CLAUDE.md instead of the project's canonical rules. This is a false positive — the file already correctly references CodeRabbit — Line 16: "Do not re-raise skipped items" wording
CodeRabbit suggests softening "Do not re-raise skipped items" to allow reopening when new changes affect context. While theoretically sound, in practice this adds ambiguity to the prompt. The current strict wording prevents the re-review step from devolving into re-litigating the same findings every iteration — which is the more likely failure mode. If a fix truly changes the context of a skipped item, the "Free review" agent (Agent 5) would catch it anyway. Current wording is better for prompt clarity. Gemini Code Assist — General review comment
Gemini's review suggested enhancing the verification step's instructions. The current step 4 is already clear: "Fix any regressions and repeat this step until all checks pass before continuing." This is sufficient and consistent with other command files in the repo. Devin — No issues foundAgreed. Clean PR. Verdict (unchanged)Approve — Clean, well-scoped PR. No new concerns. 🤖 Generated with Claude Code |
…ories Reorganize review agents to match established patterns: - Correctness → Code quality (adds code smells) - Free review → Test coverage (dedicated agent) - Conventions no longer includes test coverage - Add double-filter: agents report only noteworthy findings, triage step re-filters before classifying Fix/Skip Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add Agent 6 for overall design concerns, side effects, and integration risks that the other 5 specialized agents may miss. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
PR Review: No new feedbackPrevious Claude reviews (2 rounds) already covered this PR comprehensively. The code (single 20-line markdown file) hasn't changed since. All bot inline comments were evaluated and triaged. Verdict: Approve — Clean, well-scoped addition following established 🤖 Generated with Claude Code |
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #1285 +/- ##
=======================================
Coverage 87.18% 87.18%
=======================================
Files 115 115
Lines 4324 4324
Branches 1002 1002
=======================================
Hits 3770 3770
Misses 554 554 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
1 participant
Add a
review-loopcustom command that performs iterative code review and fix cycles.The command spawns 5 parallel review agents, each examining the diff from a different angle:
It then triages findings (fix vs skip), applies fixes, verifies with lint+test, and repeats until clean (max 3 iterations).
Checklist
npm run testnpm run lint