agentleFS
paoloirrerapaoloirrera/ccgraph/.claude/skills/pr-draft/SKILL.md

pr-draft

Skill1 starsChanged 3 months agoMIT-0View on GitHub

Draft a PR description (pr.md) from the story, design doc, tasks, and the actual changes on the branch

5 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions5 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 5

    1. forcedotcom/salesforcedx-vscode★ 1ksame name
    2. ahmadammarm/musyskills★ 1same name
    3. jis3r/skills★ 1same name
    4. omson-datco/claude-skills★ 0same name
    5. Srltas/claude-skills★ 0same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository8 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

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

  • 0

    Common

    Most of the 5 edited versions we read keep these.

  • 2

    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
  1. Very specific:

    Read story.md (scope, DoD) and design-doc.md (architecture, ADRs) if present.

    line 17
  2. Very specific:

    Read tasks/README.md to see what was in scope and each task's final status.

    line 18
  3. Distinctive:

    Reconcile against the DoD: Check each story Definition-of-Done checkbox against the diff — grep for the function, flag, or test that …

    line 27
  4. Distinctive:

    Keep it reviewer-oriented: what to look at, what to trust, what's deferred.

    line 35

“Common” is measured: kept by at least 6 in 10 of the 5 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

/pr-draft — draft the PR

Final step of the AI-DLC pipeline (after the building phase). Synthesizes a pr.md from the inception artifacts and the actual work done on the branch, ready to paste into a pull request.

Input

The story slug, e.g. /pr-draft weather-tool. Reads .local/user-stories/<slug>/{story.md, design-doc.md, tasks/} where present. If the slug is omitted, infer it from the branch and the commits on it, and say which story you inferred; if that's ambiguous, list the candidates under .local/user-stories/ and ask.

Steps

  1. Gather context:
    • Read story.md (scope, DoD) and design-doc.md (architecture, ADRs) if present.
    • Read tasks/README.md to see what was in scope and each task's final status.
    • Inspect the real diff: git status, git diff main...HEAD (or the appropriate base), and git log on the branch. The PR describes what actually changed, not just what was planned — reconcile the two and note any divergence.
  2. Read the template in references/pr-template.md (bundled).
  3. Draft pr.md:
    • Title — imperative, matches the story's intent and the repo's commit style.
    • Summary — what this PR adds/changes and why it matters, in a few sentences.
    • What's included — the concrete deliverables (files/modules/ADRs), grouped logically.
    • Testing — the exact commands a reviewer runs to verify, and what passes. Run them before writing the numbers down; never carry counts over from a previous draft or from the conversation.
    • Notes for reviewers — non-obvious decisions, intentional trade-offs, follow-ups deferred to a later story.
  4. Reconcile against the DoD. Check each story Definition-of-Done checkbox against the diff — grep for the function, flag, or test that satisfies it rather than trusting the task statuses, which record intent. If any isn't met, call it out explicitly rather than implying completeness.
  5. Show the draft and confirm before writing.
  6. On confirmation, write .local/user-stories/<slug>/pr.md. Offer to open the PR with gh only if the user asks — do not push or create the PR unprompted. (If the user is ready to open a PR, the repo convention is to run /code-review first — remind them.)

Principles

  • The PR reflects reality. If the build diverged from the design, the PR documents the as-built state and the reason, not the original plan. A divergence the reviewer would want to know about belongs in Notes, with the evidence that drove it.
  • Testing steps must be copy-pasteable and verified to pass as written. ccgraph is a single binary crate: run cargo from the crate root (ccgraph/), and the full gate is fmt --check / clippy --all-targets -- -D warnings / test per CLAUDE.md.
  • Keep it reviewer-oriented: what to look at, what to trust, what's deferred. Prefer a measured number over an adjective — "57 signals → 8 across 16 sessions" tells a reviewer where to look; "much less noisy" doesn't.

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.