Skip to content
Get packs

PR Gardening

Finds your instance's open pull requests, checks which are merge-ready, and drives the rest toward green without merging.

Source Download skill
Repository
paperclipai/paperclip
Stars
110,078
Licence
MIT
Files
6 files

What it does

This skill teaches an AI agent to review the pull requests that a Paperclip instance itself opened, based on recent Paperclip issues that mention them. Scripts (not the model) list candidate PRs, filter out community and dependabot PRs by author, and check each PR's current head for failing checks, conflicts, and review status. For PRs that are not ready, the agent runs a separate prepare-PR skill to push them toward green, while the skill itself is forbidden from merging, approving, or closing anything.

Great for

  • Maintainers of a Paperclip instance who want a regular sweep of their own open pull requests
  • Teams using Paperclip issues to track PR work who want readiness checks tied to those issues
  • Developers who want a scripted, read-only readiness report before deciding what to merge

Skip it if

  • Reviewing community or external contributors' PRs, which the skill excludes by design and forbids touching (they are out of scope)
  • Anyone outside a Paperclip deployment, since it depends on Paperclip's API and issue data; it is not a general PR tool
  • Unattended runs without a dry-run first, because it can post comments, create follow-up issues and archive inbox entries when not in dry-run mode

Try asking

  • “Run PR gardening for the last 14 days in dry-run mode and show me which of our PRs are not ready and why”
  • “Check the readiness of our open PRs and report which ones have failing Greptile checks or merge conflicts”
  • “Garden our non-ready PRs with the prepare-PR skill, respecting the 48-hour cooldown”
  • “List PRs referenced by active Paperclip issues that were dropped as stale or community contributions”
Step-by-step guide: how to use PR GardeningInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 6 original files, the MIT licence and an attribution note.

    pr-gardening.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip pr-gardening.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip pr-gardening.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Always start with --dry-run to see the candidate and readiness report before letting the agent post comments or create tasks
  • Keep the default author filter; only widen it with --authors or --include-community when you explicitly intend to
  • Review the generated candidates.json and readiness.json in the run scratch directory, since readiness is computed from GitHub checks at a specific head SHA

What's inside

  • scripts/check-readiness.mjs
  • scripts/find-candidates.mjs
  • scripts/lib.mjs
  • scripts/pr-gardening.test.mjs
  • scripts/render-report.mjs
  • SKILL.md

Redistributed unmodified from paperclipai/paperclip under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.