agentleFS
jdrydnjdrydn/dotclaude/skills/pull-request/SKILL.md

pull-request

Skill0 starsChanged 56 days agoUnlicenseView on GitHub

Create a pull request for the current branch. The goal is zero unnecessary questions — infer everything possible from the repo, branch name, and commit history, then create the PR immediately. Run these in a single batch to minimise round-trips: Use this order of precedence: 1. If the branch was checked out from develop (check git log --oneline develop..HEAD succeeds and is short), target develop. 2. If main exists, target main. 3. Otherwise target master. 4. If genuinely ambiguous (e.g.…

  • Commits and pushes

53 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use issue-campaign, contracts and 9 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • samchonSkill

    issue-campaign

    Defines the default solo repository-wide issue campaign for typia: exhaustive discovery, lead-vetted issue publication, one unified CI-validated implementation pull request per cycle, solo Self-Review, and repeated rediscovery until a full clean round. Use for broad audits, many issue candidates, or repeated issue-to-pull-request campaigns unless the user explicitly requests parallel or multi-agent execution; do not use for one already-defined issue or an ordinary pull request.

    Also in 3 projects that use this

  • AnthropicSkill

    contracts

    Answer a question across a corpus of contract documents with verified citations. Use when the user asks what a contract says, which contracts have a clause, what changed between amendments, or any question that needs reading and citing across a set of contract files. The corpus must be on the local filesystem (see README).

    Also in 3 projects that use this

  • mfang0126Skill

    development

    Full-stack development workflow skill. Guides the complete cycle from understanding requirements to shipping code via PR with Preview validation.

    Also in 3 projects that use this

  • dragonflydbSkill

    benchmark

    Benchmark Dragonfly (and compare against Valkey/Redis) on local or remote cloud instances, then produce a performance + memory report with charts and a raw-data appendix. Use this whenever the user runs /benchmark, or asks to benchmark / load-test / measure throughput, QPS, latency, or memory efficiency of Dragonfly or another Redis-compatible server, to compare Dragonfly vs Valkey or Redis, or to generate a benchmark report. Trigger it even when the user just describes the goal ("see how fast Dragonfly is on this box", "compare memory per key vs Valkey") without saying the word "benchmark".

    Also in 4 projects that use this

  • VercelSkill

    gh-stack

    Manages stacked PRs and splits multi-part work into reviewable branches with gh-stack. Use for stack creation, viewing, edits, push, submit, sync, rebase, merge, or checkout; when asked to split or isolate work for review; whenever a user mentions a stack, branch layers, dependent PRs, or gh stack; or when a stack is checked out.

    Also in 3 projects that use this

  • carapace-shSkill

    git

    Use when working with Git — object model, references, index, branching, merging, rebasing, remotes, diffs, config, hooks, worktrees, submodules, and plumbing commands. Triggers on: "git", "git branch", "git commit", "git merge", "git rebase", "git checkout", "git stash", "git reset", "git submodule", "git worktree", "git blame", "git bisect", "git reflog", "git hooks", "git config".

    Also in 3 projects that use this

The file

Show its rules and text

We split this skill into the rules it gives an agent

15 rules. 0 are common, 9 are distinctive, and 6 are very specific to this project.

  • 0

    Common

    Most of the 48 edited versions we read keep these.

  • 9

    Distinctive

    This author's own choices. Worth a look.

  • 6

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Create a pull request for the current branch.

    line 12
  2. Distinctive:

    Run these in a single batch to minimise round-trips:

    line 17
  3. Distinctive:

    Use this order of precedence:

    line 55
  4. Very specific:

    If the branch was checked out from develop (check git log --oneline develop..HEAD succeeds and is short), target develop.

    line 57
  5. Distinctive:

    If main exists, target main.

    line 59
  6. Very specific:

    If genuinely ambiguous (e.g.

    line 61
  7. Very specific:

    Look for a test command in this order: package.json scripts (test, test:unit), Makefile targets, a pytest / go test / cargo test …

    line 77
  8. Very specific:

    If the repo uses Conventional Commits (check the last ~20 commits on the target branch for feat:, fix:, chore: patterns), match that style.

    line 85
  9. Distinctive:

    Keep it under 72 characters.

    line 89
  10. Distinctive:

    Follow its structure exactly.

    line 95
  11. Distinctive:

    Before asking the user, try to find a reference automatically:

    line 127
  12. Very specific:

    If found, include it in the body (in the Why section or as a Closes #N / Refs: PROJ-123 footer).

    line 133
  13. Distinctive:

    Do this immediately. Don't ask the user to review the title or body — they invoked this skill because they want a PR, not a preview.

    line 139
  14. Very specific:

    If the user said "draft" or the branch name starts with draft/ or wip/, add --draft.

    line 149
  15. Distinctive:

    Never fabricate test results: If tests weren't run, say so.

    line 163

“Common” is measured: kept by at least 6 in 10 of the 48 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

Pull Request

Create a pull request for the current branch. The goal is zero unnecessary questions — infer everything possible from the repo, branch name, and commit history, then create the PR immediately.

1. Gather context (no user interaction needed)

Run these in a single batch to minimise round-trips:

# Current branch and remote
git rev-parse --abbrev-ref HEAD
git remote get-url origin 2>/dev/null

# Detect target branch (first match wins)
git rev-parse --verify develop 2>/dev/null && echo "TARGET=develop" \
  || git rev-parse --verify main 2>/dev/null && echo "TARGET=main" \
  || echo "TARGET=master"

# Commits in this PR
git log <target>..HEAD --oneline

# Diff stats
git diff --stat <target>..HEAD
git diff --stat <target>..HEAD | tail -1   # summary line

# Behind-target check
git fetch --quiet
git log HEAD..<target> --oneline

# Uncommitted changes
git status --porcelain

# PR template
cat .github/pull_request_template.md 2>/dev/null \
  || cat .github/PULL_REQUEST_TEMPLATE.md 2>/dev/null \
  || cat .github/PULL_REQUEST_TEMPLATE/*.md 2>/dev/null \
  || echo "NO_TEMPLATE"

# Work plan (if it exists)
ls docs/work/*/plan.md 2>/dev/null

Target branch heuristic

Use this order of precedence:

  1. If the branch was checked out from develop (check git log --oneline develop..HEAD succeeds and is short), target develop.
  2. If main exists, target main.
  3. Otherwise target master.
  4. If genuinely ambiguous (e.g. both develop and main are plausible), ask.

2. Pre-flight guards

Evaluate these before writing any description. Act on the first that fires.

ConditionAction
Test suite is failingStop. Tell the user which tests failed and refuse to create the PR unless they explicitly override with "skip tests" or similar.
Uncommitted changes presentWarn. Suggest committing or stashing. Proceed only if the user acknowledges.
Branch is behind targetWarn. Show how many commits behind. Suggest rebasing. Proceed if the user acknowledges.
Diff exceeds ~500 added/removed linesWarn. Note the size and suggest splitting if logical boundaries exist. Still proceed — large PRs are sometimes unavoidable.
No commits ahead of targetStop. There's nothing to open a PR for.

Running tests

Look for a test command in this order: package.json scripts (test, test:unit), Makefile targets, a pytest / go test / cargo test invocation, or a CI config that reveals the command. If nothing obvious exists, skip the test step and note that no test suite was detected.

3. Build the PR title

Derive the title from the branch name and commit history — don't ask.

  • If the repo uses Conventional Commits (check the last ~20 commits on the target branch for feat:, fix:, chore: patterns), match that style. Derive the type and scope from the branch name or commits: feature/user-auth → feat(auth): add user authentication
  • Otherwise, use imperative mood, sentence case: bugfix/broken-login → Fix broken login flow
  • Keep it under 72 characters.

4. Build the PR body

If a PR template exists

Follow its structure exactly. Fill in every section using the commit log and diff. Leave nothing as a placeholder — write real content or remove the section if it genuinely doesn't apply.

If no template exists

Use this structure:

## What

<Summarise what changed, grouped by concern — not a raw file list.
 2–5 bullet points for a typical PR. For single-commit PRs, a short
 paragraph is fine.>

## Why

<Motivation and context. What problem does this solve? Link to any
 issue/ticket detected in step 5 below.>

## How to test

<Concrete verification steps. "Run the test suite" is acceptable if
 there are good automated tests; otherwise describe manual steps.>

## Notes

<Breaking changes, migration steps, deployment considerations.
 Omit this section entirely if there's nothing to flag.>

5. Auto-detect issue/ticket references

Before asking the user, try to find a reference automatically:

  1. Branch name — patterns like PROJ-123, #45, GH-12, or issue-99 embedded in the branch name.
  2. Commit messages — scan for the same patterns, plus Fixes #N, Closes #N, Relates to PROJ-NNN.
  3. Work plan — if docs/work/<slug>/plan.md exists, scan it for ticket/issue references.

If found, include it in the body (in the Why section or as a Closes #N / Refs: PROJ-123 footer). If nothing is found, don't ask — just skip it. The user can always add it later; the friction of being asked every time is worse than occasionally missing a link.

6. Create the PR

Do this immediately. Don't ask the user to review the title or body — they invoked this skill because they want a PR, not a preview.

gh pr create \
  --title "..." \
  --body "..." \
  --base <target>

If the user said "draft" or the branch name starts with draft/ or wip/, add --draft.

7. Present the result

Show:

  • The PR URL (clickable).
  • A one-line summary of what was opened (title, target branch, number of commits, rough line count).

That's it — no lengthy recap of the description.

Hard rules

  • Never mention Claude, AI, or any AI tool in the PR title or body.
  • Never fabricate test results. If tests weren't run, say so.
  • Never include raw file lists as the primary description — always summarise by concern or area.

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.