pr-shepherding
Scannednpx machina-cli add skill a5c-ai/babysitter/pr-shepherding --openclawPR Shepherding
Overview
Monitor and manage a PR from creation through merge. Handle CI failures, respond to review comments, resolve threads, and verify merge readiness.
When to Use
- After PR creation from completed work
- When a PR needs CI monitoring
- When review comments need automated handling
Process
- Monitor CI - Watch pipeline status, triage failures
- Handle Comments - Respond to review feedback, make code changes
- Resolve Threads - Close addressed review threads
- Verify Readiness - Check approvals, CI green, threads resolved, coverage met
Merge Readiness Checklist (gtg)
- All CI checks passing
- Required approvals obtained
- All review threads resolved
- Coverage thresholds met
- No merge conflicts
- SERVICE-INVENTORY.md updated
Agents Used
agents/pr-shepherd/- PR lifecycle management
Tool Use
Invoke via babysitter process: methodologies/metaswarm/metaswarm-pr-shepherd
Source
git clone https://github.com/a5c-ai/babysitter/blob/main/plugins/babysitter/skills/babysit/process/methodologies/metaswarm/skills/pr-shepherding/SKILL.mdView on GitHub Overview
PR Shepherding ensures end-to-end oversight of a PR from creation through merge. It automates CI monitoring, responds to review comments, resolves threads, and verifies merge readiness.
How This Skill Works
The skill continuously monitors pipeline status and triages failures, then automates responses to review feedback and applies necessary code changes. It also resolves addressed threads and runs a merge-readiness verification to ensure approvals, green CI, resolved threads, adequate coverage, and no conflicts before merging.
When to Use It
- After PR creation from completed work
- When a PR requires CI monitoring
- When review comments need automated handling
- When reviewers have numerous threads that require resolution
- When ensuring merge readiness (approvals, CI green, coverage, conflicts, inventory update)
Quick Start
- Step 1: Invoke via babysitter process: methodologies/metaswarm/metaswarm-pr-shepherd
- Step 2: Monitor CI, triage failures, and respond to review comments automatically
- Step 3: Run the merge-readiness verification and ensure all checklist items are satisfied
Best Practices
- Define clear CI gates and monitor pipeline status continuously
- Triage CI failures promptly and document actionable fixes
- Automate responses to common review comments with targeted changes
- Proactively resolve threads to close discussions and prevent drift
- Run the merge-readiness checklist and ensure SERVICE-INVENTORY.md is updated before merge
Example Use Cases
- A PR triggers a failing CI pipeline; pr-shepherd reruns tests, comments with suggested fixes, and applies changes automatically.
- Automated replies address reviewer feedback and push patches, reducing back-and-forth.
- All relevant threads are resolved after changes, marking the PR as ready for final review.
- Merge readiness is tested by confirming approvals, green CI, and no conflicts before merging.
- SERVICE-INVENTORY.md is updated as part of the readiness verification to reflect changes in the PR.