modules/programs/agents/shared/skills/shepherd-pr/SKILL.md
Shepherd an already-open PR across the line: monitor CI + bot comments, invoke `fix-pr` on issues, then merge when clean (respecting repo branch protection). Composes `monitor-pr` + `merge-pr`. Use when user says "shepherd", "shepherd PR", "watch and merge", or wants hands-off ship-it behavior on an existing PR.
npx skillsauth add MichaelVessia/nixos-config shepherd-prInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Arguments: $ARGUMENTS (optional PR number or URL; else current branch PR).
This skill is a recipe: monitor-pr → merge-pr. All real work is delegated.
Invoke monitor-pr with $ARGUMENTS.
Invoke merge-pr with $ARGUMENTS.
merge-pr enforces all policy (no CHANGES_REQUESTED bypass, uses repo's
allowed merge method, uses gh pr merge --auto to respect branch
protection).
Combine the two sub-reports into a single final summary:
## Shepherd-PR
- **PR**: <URL>
- **Monitor outcome**: clean | blocked (<reason>)
- **Merge outcome**: merged | queued (auto-merge) | not attempted | blocked (<reason>)
- **fix-pr invocations**: 0 | 1 | 2 | 3
If the PR was queued via auto-merge, tell the user it will land on its own once branch protection clears.
development
Generate self-contained HTML visualizations with Plannotator theming. Use for implementation plans, PR explainers, architecture diagrams, data tables, slide decks, and any visual explanation of technical concepts. Plans and PR explainers follow Plannotator's prescriptive approach; all other visual content delegates to nicobailon/visual-explainer.
development
Turn an idea or objective into a goal package for /goal. Interviews the user, builds a reviewed fact sheet via Plannotator, then explores the codebase to produce an execution plan.
development
Open Plannotator's browser-based code review UI for the current worktree or a pull request URL, then act on the feedback that comes back.
testing
Open Plannotator on the latest rendered assistant message and use the returned annotations to revise that message or continue.