skills.rest
mslshao
mslshaoCOMMUNITY
Explore all skills by this author →
Website𝕏@mslshaoGitHub

babysit-pr

Poll draft GitHub pull requests and auto-remediate mechanical bot feedback.

1|Updated May 15, 2026
Search Tags:#automation#code review#pull request#github#worktree#force push#ci monitoring
One-click install
npx skills add https://github.com/mslshao/claude-harness --skill babysit-pr-mslshao
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: babysit-pr
Source: https://github.com/mslshao/claude-harness/tree/main/dotclaude/skills/babysit-pr
Command: npx skills add https://github.com/mslshao/claude-harness --skill babysit-pr-mslshao

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill reduces the overhead of babysitting an open draft pull request when automated reviewers, CI systems, and bots keep posting updates while you are away. It helps you avoid losing momentum by continuously watching PR activity, handling safe mechanical fixes, and pausing only when real human judgment is required.

Core Features & Use Cases

  • Autonomous PR polling loop: Re-checks a draft PR on a paced schedule, persists state across wakeups, and resumes cleanly even after session compaction.
  • Smart comment classification: Separates bot noise, mechanical inline suggestions, review-state changes, and human reviewer feedback into ignore, reply, auto-remediate, or escalate actions.
  • Safe auto-remediation workflow: Uses a dedicated worktree, local validation checks, amend-and-force-push flow, and strict authorization gates before applying small mechanical fixes.
  • Durable audit trail: Logs every cycle, action, escalation, and PR snapshot into a tracking bead so the operator can review exactly what happened.
  • Use Case: Start this when you open a draft PR and expect Copilot, CI bots, or code-quality bots to comment while you step away, so the loop can absorb trivial feedback and alert you only when a human reviewer or substantive issue appears.

Quick Start

Ask the AI to babysit your draft PR by number and include force-push authorization only if you want it to auto-apply small mechanical bot fixes.

Frequently Asked Questions about babysit-pr

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate GitHub draft pull request monitoring for bot comments?▼

Automate GitHub draft pull request monitoring by running a polling loop that checks comments, reviews, and CI state on a paced schedule. The loop classifies bot noise, auto-remediates trivial inline suggestions in a worktree, and escalates human feedback.

Can I automatically apply code review suggestions from bots to my pull request?▼

Yes, you can automatically apply mechanical code review suggestions from bots. The Skill uses a dedicated worktree, runs local validation checks, and performs an amend-and-force-push flow under strict authorization gates to safely apply small fixes.

How do I handle CI status changes and bot noise on a pull request while away?▼

Handle CI status changes and bot noise while away by persisting state across scheduled wakeups or cron fallbacks. The Skill resumes cleanly after session compaction, logs every cycle into a tracking bead, and pauses only when real human judgment is required.

Do I need GitHub CLI to run an automated pull request babysitting loop?▼

Yes, GitHub CLI access is required to run the automated pull request babysitting loop. You also need scheduled wakeups or a cron fallback for the polling schedule, durable state tracking for resuming sessions, and worktree isolation for applying fixes.

What is the best way to separate human reviewer feedback from bot noise on GitHub?▼

The best way to separate human reviewer feedback from bot noise is smart comment classification. The Skill categorizes incoming PR activity into ignore, reply, auto-remediate, or escalate actions based on whether the source is mechanical or requires human judgment.

When should I not use an automated pull request auto-remediation workflow?▼

You should not use an automated pull request auto-remediation workflow when a human reviewer posts substantive feedback or complex issues. The Skill escalates these cases and pauses, applying amend-and-force-push only for small mechanical fixes after local validation passes.

Related Skills

Explore More Software Engineering Skills→
kanade0404kanade0404

pr-monitor

Monitor GitHub pull requests until merged or closed, then trigger the retro workflow.

Community
Advanced
sensei-hackersensei-hacker
4

pr-review

Review GitHub pull requests, bot comments, and CI checks.

Community
Intermediate
AzrinPutraAzrinPutra

code-review

Review code changes, auto-fix mechanical issues, and suggest improvements.

Community
Advanced
matt-rileymatt-riley
1

review-comment-resolution

Classify PR review comments and apply validated fixes to code branches.

Community
Intermediate