babysit-pr

Monitor pull requests through review and CI until checks pass.

3|1|Updated Aug 23, 2025
One-click install
npx skills add https://github.com/samuelpatro/.claude --skill babysit-pr-samuelpatro
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: babysit-pr
Source: https://github.com/samuelpatro/.claude/tree/main/skills/babysit-pr
Command: npx skills add https://github.com/samuelpatro/.claude --skill babysit-pr-samuelpatro

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Keeping a pull request moving through AI review bots and CI requires constant attention: verifying bot claims, replying to threads, rerunning flaky checks, and knowing when to stop. This Skill automates that shepherding loop so the PR reaches green without spamming reviewers or expanding scope. ## Core Features & Use Cases - Bot Review Triage: Treats every AI review bot comment as a hypothesis, verifies it against the actual code, fixes real bugs, and pushes back with evidence on wrong ones. - CI Failure Handling: Reads failing logs with gh run view --log-failed, fixes what the PR caused, and reruns flaky checks once before declaring them real. - Thread Management: Replies to and resolves inline review threads via the GitHub GraphQL API, respecting that human reviewers outrank bots. - Use Case: You open a PR and CodeRabbit leaves eight comments while two CI jobs fail. Ask the agent to babysit the PR: it verifies each comment, fixes the two real issues, replies with file:line evidence on the wrong ones, reruns the flaky job, and reports when everything is green. ## Quick Start Ask the agent to babysit the current branch's pull request and handle the review comments until CI is green.

Frequently Asked Questions about babysit-pr

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

FAQPage Schema
How do I handle CodeRabbit review comments on a pull request?▼

Treat each CodeRabbit comment as a hypothesis and read the code it points at before acting. Fix real bugs, reply with file:line evidence when the claim is wrong, and resolve the thread via the GitHub GraphQL API.

How do I reply to an inline PR review comment with the GitHub CLI?▼

Inline replies require the GraphQL API, since gh pr comment only posts top-level comments. Use gh api graphql with the addPullRequestReviewThreadReply mutation, passing the thread ID and reply body.

Why does CodeRabbit skip reviewing my draft pull request?▼

Review bots skip draft PRs, posting a 'Review skipped: Draft detected' note. Either mark the PR ready with gh pr ready, or trigger one manual review with a single @coderabbitai review comment and wait.

When should I rerun a failing CI check on a pull request?▼

Rerun a flaky check once with gh run rerun --failed after reading the failing log. If it fails again, treat it as real; if it also fails on the base branch, report it instead of fixing it in this PR.

What are the limits of automated PR babysitting?▼

It only acts on checks and comments newer than the latest push, never resolves human reviewers' threads, and avoids scope creep beyond the PR's original goal. Merging happens only when the user explicitly requests it.