pull-request-feedback-loop
Pull Request Feedback Loop
PR feedback is a queue of review obligations, not a pile of comments. Use this skill when a user asks you to address PR comments, requested changes, failed checks tied to review feedback, or stale review threads.
When to Use
- A PR has unresolved review comments or requested changes
- The user asks you to "address review feedback" or "fix PR comments"
- You need to reconcile code changes, replies, and verification
Do not use this for ordinary local refactors with no review thread.
Process
-
Identify the PR.
- Prefer the PR number or URL from the user.
- If missing, run
gh pr statusorgh pr view --json number,url,headRefName.
-
Fetch review context.
- Run
gh pr view <pr> --comments. - If needed, run
gh apifor review threads or inline comments. - Capture author, file, line, status, and requested change.
- Run
-
Group comments by issue.
- Merge duplicate comments that point at the same underlying fix.
- Separate behavior bugs, test gaps, naming/docs cleanup, and questions.
- Mark comments that need clarification before code changes.
-
Fix one group at a time.
- Inspect the referenced code before editing.
- Make the smallest coherent change for that group.
- Avoid bundling unrelated cleanup into review fixes.
-
Verify before replying.
- Run the narrowest relevant test or check for each group.
- If no test can run, explain the exact reason and inspect manually.
- Do not claim a thread is resolved until verification passes or is explicitly waived.
-
Prepare replies.
- Summarize what changed and how it was verified.
- Keep replies short and factual.
- Do not submit replies automatically unless the user asked you to.
Output
Report:
- Groups handled
- Files changed
- Verification run
- Threads ready for reply
- Threads needing user or reviewer clarification
Guardrails
- Do not force-push, delete branches, or dismiss reviews without explicit permission.
- Do not mark comments resolved only because code was edited.
- Do not ignore a comment because it looks small; either address it or explain why it is not applicable.
More from archieindian/openclaw-superpowers
heartbeat-governor
Enforces per-skill execution budgets for scheduled cron skills — pauses runaway skills that exceed their token or wall-clock budget before they drain your monthly API allowance.
25context-window-management
Prevents context overflow on long-running OpenClaw sessions. Use when approaching context limits.
25fact-check-before-trust
Triggers a secondary verification pass for any agent output containing factual claims, numbers, dates, or named entities before the output is acted on
19long-running-task-management
Breaks multi-hour tasks into checkpointed stages with resume capability. Use when a task is expected to take more than 30 minutes or multiple sessions.
19agent-self-recovery
Detects when the agent is stuck in a loop and escapes systematically. Use when you notice repeated failures or loss of direction.
18skill-vetting
Reviews a ClawHub skill's source code for security risks before installation. Use before installing any new skill.
18