pr-review-handler
Skill0 starsChanged 9 months agoMITView on GitHub
Handle PR review feedback by checking for comments, determining necessary changes, responding appropriately to reviewers, implementing fixes, and pushing updates. Use when the user asks to work on an existing open PR, address PR feedback, fix review comments, or handle bot review suggestions (Devin AI, Cursor Bugbot, etc.).
- Commits and pushes
We split this skill into the rules it gives an agent
19 rules. About 0 are common, 16 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈16
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 12
Use this skill when:
- Distinctive: line 58
Create a prioritized list of changes:
- Distinctive: line 80
Verify the fix (run tests, build, etc.)
- Distinctive: line 86
Use clear commit messages: "Fix: <what> per <reviewer> feedback"
- Distinctive: line 115
Ask follow-up questions if unclear
- Distinctive: line 174
When reviewer requests significant refactoring:
- Distinctive: line 177
If in-scope: Implement and push
- Distinctive: line 178
If out-of-scope: Respond proposing separate PR/issue
- Distinctive: line 179
If unclear: Ask reviewer to clarify scope expectations
- Distinctive: line 210
Read before responding: Always fetch and read all comments before starting fixes
- Distinctive: line 214
Verify before pushing: Run tests/build to avoid pushing broken code
- Distinctive: line 215
Update PR description: If changes significantly alter the PR, update description
- Very specific: line 216
Re-request review: Use gh pr ready <number> if PR was marked as draft
- Very specific: line 221
Check if gh is installed: gh --version
- Very specific: line 222
Check if authenticated: gh auth status
- Distinctive: line 223
Check if PR number is valid
- Distinctive: line 226
Check for conflicts: git status
- Distinctive: line 232
Ask reviewer for guidance
- Distinctive: line 233
Don't leave comment unanswered
≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
PR Review Handler
Handle pull request review feedback systematically: fetch comments, analyze feedback, implement fixes, respond to reviewers, and push updates.
When to Use
Use this skill when:
- User asks to "work on PR #X" or "address PR feedback"
- User provides a PR URL and asks to fix review comments
- User asks to "respond to review comments" or "fix the PR issues"
- Bot reviewers (Devin AI, Cursor Bugbot) have left feedback
- Multiple reviewers have left comments requiring coordination
Workflow
1. Fetch PR Information
Get the PR number from the user or URL, then fetch details:
# Get PR metadata
gh pr view <number> --json number,title,state,baseRefName,headRefName,author,url
# Get reviews and their state
gh pr view <number> --json reviews --jq '.reviews[] | "Review by \(.author.login) - \(.state):\n\(.body)\n---"'
# Get review comments (line-specific feedback)
gh api repos/<owner>/<repo>/pulls/<number>/comments --jq '.[] | "File: \(.path)\nLine: \(.line // .original_line)\nComment by \(.user.login):\n\(.body)\n---"'2. Analyze Feedback
For each comment, categorize by:
Priority:
- Critical: Security issues, bugs that cause crashes, data loss
- High: Bugs, performance issues, incorrect implementations
- Medium: Code quality, refactoring suggestions, best practices
- Low: Style issues, naming suggestions, optional improvements
Actionability:
- Actionable: Clear change requested with specific fix
- Question: Reviewer asking for clarification
- Discussion: Design discussion or trade-off debate
- Informational: FYI comment, no action needed
Bot vs Human:
- Bot comments (Devin AI, Cursor Bugbot): Often auto-fixable, may have suggested code
- Human comments: May require judgment, discussion, or clarification
3. Plan Changes
Create a prioritized list of changes:
- Group related comments (e.g., all comments about validation)
- Identify dependencies (Fix A must happen before Fix B)
- Separate code changes from response-only comments
- Note which comments need discussion vs implementation
4. Checkout Branch
git fetch origin
git checkout <head-branch-name>
git pull origin <head-branch-name>5. Implement Fixes
Work through changes by priority:
For each fix:
- Read relevant files
- Implement the change
- Verify the fix (run tests, build, etc.)
- Stage changes:
git add <files>
Commit strategy:
- Single fix commit: If all changes are tightly related
- Multiple commits: If addressing distinct concerns (validation, refactoring, bug fix)
- Use clear commit messages: "Fix:
<what>per<reviewer>feedback"
6. Respond to Comments
Respond to each comment appropriately:
For implemented fixes:
gh api -X POST repos/<owner>/<repo>/pulls/<number>/comments/<comment-id>/replies \
-f body="Fixed in <commit-sha>. <Brief explanation of fix>"For questions/clarifications:
gh api -X POST repos/<owner>/<repo>/pulls/<number>/comments/<comment-id>/replies \
-f body="<Clear answer to question>"For won't-fix with reasoning:
gh api -X POST repos/<owner>/<repo>/pulls/<number>/comments/<comment-id>/replies \
-f body="<Explanation of why not changing, with reasoning>"Response guidelines:
- Be concise: 1-3 sentences
- Reference commit SHAs for fixes
- Explain reasoning for design decisions
- Thank reviewers for catching issues
- Ask follow-up questions if unclear
7. Push Changes
git push origin <head-branch-name>8. Verify and Summarize
After pushing:
# Check CI status
gh pr checks <number>
# View updated PR
gh pr view <number>Provide user with summary:
- Number of comments addressed
- Commits pushed
- Comments that need discussion
- Any blockers or questions
Common Patterns
Pattern 1: Bot Review Fixes
Bot reviews (Devin AI, Cursor Bugbot) often suggest specific code:
- Fetch comments
- For each bot suggestion:
- Assess validity (is this actually an issue?)
- Implement fix if valid
- Respond with "Fixed in
<sha>" or explain why not applicable
- Group related bot fixes into single commit
- Push once after addressing all bot feedback
Pattern 2: Mixed Human + Bot Reviews
- Address critical human feedback first
- Handle bot feedback second
- Respond to human questions/discussions
- Push changes
- Respond to all comments with commit references
Pattern 3: Conflicting Feedback
When reviewers disagree:
- Acknowledge both perspectives in responses
- Make a reasoned decision based on project context
- Explain decision to both reviewers
- Offer to discuss synchronously if needed
Pattern 4: Large Refactoring Requests
When reviewer requests significant refactoring:
- Assess scope (does this belong in this PR?)
- If in-scope: Implement and push
- If out-of-scope: Respond proposing separate PR/issue
- If unclear: Ask reviewer to clarify scope expectations
Response Templates
Fixed:
Fixed in <sha>. <One sentence explaining what changed>Won't fix with reason:
Keeping as-is because <reasoning>. <Alternative if applicable>Clarification question:
<Answer to question>. <Additional context if helpful>Agreement to refactor separately:
Good point. I'll address this in a follow-up PR to keep this change focused on <original scope>.Design decision explanation:
I chose <approach> over <alternative> because <reasoning>. Open to changing if you feel strongly.Best Practices
- Read before responding: Always fetch and read all comments before starting fixes
- Commit atomically: Each commit should address a coherent set of changes
- Respond to everything: Even "Fixed in
<sha>" is better than silence - Be professional: Thank reviewers, explain decisions, admit mistakes
- Verify before pushing: Run tests/build to avoid pushing broken code
- Update PR description: If changes significantly alter the PR, update description
- Re-request review: Use
gh pr ready <number>if PR was marked as draft
Error Handling
If gh command fails:
- Check if gh is installed:
gh --version - Check if authenticated:
gh auth status - Check if PR number is valid
If git push fails:
- Check for conflicts:
git status - Pull latest:
git pull origin <branch> --rebase - Resolve conflicts if needed
If unable to implement fix:
- Respond to comment explaining blocker
- Ask reviewer for guidance
- Don't leave comment unanswered
Examples
See references/examples.md for complete examples of handling different review scenarios.
In this repository29 other files it gives its agents.
- .agents/skills/vercel-composition-patterns/AGENTS.mdAGENTS.md
- .agents/skills/vercel-react-best-practices/AGENTS.mdAGENTS.md
- agent-browserSkill
- ai-sdkSkill
- algorithmic-artSkill
- audit-websiteSkill
- before-and-afterSkill
- core-web-vitalsSkill
- find-skillsSkill
- frontend-designSkill
- mcp-builderSkill
- next-best-practicesSkill
- pdfSkill
- performanceSkill
- prdSkill
- programmatic-seoSkill
- ralphSkill
- seo-auditSkill
- skill-creatorSkill
- theme-factorySkill
- ui-ux-pro-maxSkill
- using-git-worktreesSkill
- vercel-composition-patternsSkill
- vercel-react-best-practicesSkill
- webapp-testingSkill
- web-design-guidelinesSkill
- cold-outreachSkill
- continuous-claudemd-updatesSkill
- design-tasteSkill
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
Reviews can't be read right now.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

