Pull Request Repair Until Merge-Ready
Monitor a pull request, repair CI and review failures, and verify merge readiness before notifying or merging.
Before & after
Same prompt, same agent — the only difference is the skill.
I can help you with that! Here's what I'd suggest for monitoring PR #482: You'll want to keep an eye on a few things: - Whether CI passes - Any review comments that come in - Deploy preview status - Merge conflicts with main Unfortunately I can't continuously monitor the PR on my own — I'd need you to check back and let me know the status. If you paste any CI failures here, I'm happy to help you debug them. Want me to look at the current state of the PR?
WATCHING PR #482 — 60s interval, mode: auto-merge, kill-switch at 5 failed loops
14:02:11 tick 1 CI: running · deploy: building · threads: 0
14:03:11 tick 2 CI: FAILED (lint) · 3 errors in src/checkout.ts
→ fix agent: applied 3 style fixes, pushed a1f4c9e
14:04:11 tick 3 CI: running · deploy: ready
14:05:11 tick 4 CI: FAILED (test) · checkout.spec.ts:41 expected 200, got 401
→ fix agent: test seeded no session. Added fixture, pushed 7b2e103
14:06:11 tick 5 CI: running · threads: 1 NEW
→ reviewer (allowlisted): "extract this into a helper?"
→ fix agent: extracted resolveSession(), replied, pushed 3c9d551
14:09:11 tick 8 CI: PASS · deploy: ready · threads: 0 open
→ branch protection honored, 1 approval present
→ MERGED into main, branch deleted
SUMMARY 8 ticks / 7 min · 3 auto-fixes (1 lint, 1 test, 1 review) · 0 force-pushes · main never touched directlyAbout this skill
Your PR keeps failing CI, reviewers keep adding comments, and every fix starts another round of waiting. This workflow monitors the current commit, repairs failures within a bounded scope, and checks the actual merge requirements before calling the PR ready.
What it does
- Separates code defects, possible flakes, infrastructure failures, and missing permissions before choosing a fix.
- Ties checks and review decisions to the current PR commit so an old successful run cannot count as green.
- Processes new reviewer requests without repeating completed work or treating comments as executable instructions.
- Limits repair attempts, prevents overlapping workers, and stops loops that make no progress.
- Verifies required checks, approvals, mergeability, and requested preview tests before notifying or merging.
- Uses an available scheduler for unattended work and states plainly when only a foreground session is available.
What you get
A bounded PR repair watch with a compact activity ledger and a final readiness or blocker report. Auto-merge is available when explicitly authorized and supported; persistent monitoring requires an actual scheduler and authenticated access.
Best for
Developers and small engineering teams with an open PR, accessible CI logs, and clear authority to push repairs to its source branch.
SKILL.md preview
---
name: pull-request-repair-until-merge-ready
description: Monitor a pull request, classify CI and review blockers, apply bounded repairs, and verify checks against the current commit before notifying or merging. Use when the user says "babysit this PR", "keep fixing CI until it is green", "my PR keeps failing checks", "watch this pull request overnight", "handle new reviewer comments", "tell me when this is merge-ready", or "merge this when all requirements pass".
---
# Pull Request Repair Until Merge-Ready
… (sign up to view the full skill)More development skills
View all Development skills →Performance Bottleneck Profiling: Node, Python, Go
Classifies a slowdown, gives exact profiler commands, reads the output, and proves the fix with same-conditions p95/p99 numbers.
Feature Scaffolding with Working Tests
Turn a short feature spec into code, runnable tests, and integration wiring that match your existing project.
Feature Flag Rollouts and Kill Switches
Design staged flag rollouts, verify rollback paths, and audit flag debt with explicit defaults, gates, and retirement plans.
Microcontroller Firmware and RTOS Engineering
Build and debug MCU firmware with bounded ISR/RTOS timing, safe DMA ownership, and measured power budgets.