agentleFS
Docat0209Docat0209/SoloForge/skills/pre-pr/SKILL.md

pre-pr

Skill1 starsChanged 7 months agoMITView on GitHub

Pull request quality gate. Triggers ONLY when: about to run gh pr create. Do NOT trigger when: discussing PR strategy or reviewing existing PRs.

  • Commits and pushes

10 other projects use a version of this.

Adopted by

Most recently taken up by

The file

Show its rules and text

We split this skill into the rules it gives an agent

15 rules. 0 are common, 13 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 8 edited versions we read keep these.

  • 13

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  • could be automatic: a setting could enforce it
  1. Distinctive:

    Before creating any pull request, run through every section below.

    line 8
  2. Distinctive:

    If adding a new API, include a usage example in the same PR (Google).

    line 15
  3. Distinctive:

    When in doubt, make the PR smaller — reviewers never complain about PRs too small (Google).

    line 16
  4. Distinctive:

    Every item must pass before creating the PR:

    line 20
  5. Distinctive:

    Run git status — confirm nothing unpushed

    line 23
  6. Distinctive:

    Never create a PR with local-only commits

    line 24
  7. Very specific:

    If targeting main without staging verification: STOP and run staging-verify first

    line 29
  8. Distinctive:

    Every PR must trace back to an issue — no orphan PRs

    line 33
  9. Distinctive:

    Before/after screenshots for visual changes

    line 49
  10. Distinctive:

    Never merge a PR with failing checks

    line 54
  11. Distinctive:

    Look for: debug code, TODO comments, unused imports, hardcoded values

    line 60
  12. Distinctive:

    Leave comments on non-obvious decisions

    line 61
  13. Very specific:

    Never merge dev → main without passing staging-verify (Playwright verification on staging environment)

    line 71
  14. Distinctive:

    Never force-push to a shared PR branch after review has started

    line 72
  15. Distinctive:

    Every PR must leave the system in a working state (Google)

    line 73

“Common” is measured: kept by at least 6 in 10 of the 8 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Pre-PR Quality Gate

Before creating any pull request, run through every section below. All items must pass.

PR Size (Google Small CLs)

  • Ideal: 50-100 lines changed. Maximum: 200 lines. Over 200 → split into stacked PRs.
  • One PR = one concern — never bundle unrelated changes (Google).
  • Separate refactoring PRs from feature/bugfix PRs — never mix formatting with logic (Google).
  • If adding a new API, include a usage example in the same PR (Google).
  • When in doubt, make the PR smaller — reviewers never complain about PRs too small (Google).

PR Checklist

Every item must pass before creating the PR:

  1. All commits pushed to remote
    • Run git status — confirm nothing unpushed
    • Never create a PR with local-only commits
  2. PR targets the correct branch
    • Feature/fix PRs target dev — never directly to main
    • Release PRs (dev → main) are allowed ONLY after staging-verify passes with all changes verified
    • If targeting main without staging verification: STOP and run staging-verify first
  3. PR links to issue
    • Include Closes #NUMBER or Fixes #NUMBER in the PR body
    • Every PR must trace back to an issue — no orphan PRs
  4. PR description follows the standard format
    ## Summary
    - What changed and why (1-3 bullet points)
    
    ## Test plan
    - How to verify this works
  5. Tests included in same PR as feature
    • Unit tests at minimum; integration and E2E where appropriate
    • No "we'll add tests later" — an untested PR is an incomplete PR
  6. UI changes: include screenshots or recordings
    • Before/after screenshots for visual changes
    • Screen recordings for interaction changes
  7. CI passing
    • Run gh pr checks NUMBER --watch --fail-fast
    • Never merge a PR with failing checks
  8. Documentation updated if the PR changes how users build/test/interact (Google)
    • README, API docs, setup guides — update in the same PR, not a follow-up
  9. Self-review: git diff dev...HEAD — read every line one more time
    • Look for: debug code, TODO comments, unused imports, hardcoded values
    • Leave comments on non-obvious decisions

PR Title

  • Conventional format: feat: add document upload or fix: sidebar collapse
  • Short, under 70 characters
  • Details go in the body, not the title

Critical Rules

  • Never merge dev → main without passing staging-verify (Playwright verification on staging environment)
  • Never force-push to a shared PR branch after review has started
  • Every PR must leave the system in a working state (Google)

Next Steps

Report to user: "PR #[N] created: [title]. Target: [branch]. Size: [+N/-N lines]"

Suggested next steps (user decides):

  • UI changes → "Run ux-audit before merge"
  • Backend only → "Wait for CI, then merge"
  • Targets main → "Run staging-verify first"

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.