agentleFS
glennawatsonglennawatson/CSharpAgentSkills/skills/commit-messages/SKILL.md

commit-messages

Skill2 starsChanged 42 days agoMITView on GitHub

Use when writing a git commit message. Produces a Conventional Commits subject plus a short bulleted body — concise, simple sentences, one intent per commit. No novels, no walls of prose.

8 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

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

  • 0

    Common

    Most of the 6 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 4

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    No trailing period. Aim for ≤ 50 chars, hard cap ~72.

    line 25
  2. Very specific:

    Use - bullets. One change or reason per bullet.

    line 45
  3. Very specific:

    Add ! after the type/scope: feat(api)!: drop v1 token format.

    line 53
  4. Distinctive:

    Default to one logical thing per commit.

    line 58
  5. Very specific:

    If the message needs multiple feat:/fix: ideas, split into multiple commits.

    line 59
  6. Very specific:

    Don't narrate the testing regime in gory detail.

    line 66

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

Commit messages

A commit message explains what changed and why in the fewest words that are still meaningful. One commit = one intent. If you can't describe it without "and", it's probably two commits.

Format

Follow Conventional Commits:

<type>(<optional scope>): <subject>

- <bullet describing one change>
- <bullet describing another change>

<optional footer>

Subject line — type(scope): subject

  • Lowercase, imperative mood ("add", "fix", "remove" — not "added"/"adds").
  • No trailing period. Aim for ≤ 50 chars, hard cap ~72.
  • scope is optional and is the area touched, e.g. feat(auth): ....

Types:

TypeUse for
feata new feature (user-visible capability)
fixa bug fix
docsdocumentation only
refactorcode change that neither fixes a bug nor adds a feature
perfa performance improvement
testadding or fixing tests
buildbuild system, dependencies
ciCI configuration
choremaintenance, no src/test change
styleformatting/whitespace, no behavior change

Body — bullets, not paragraphs

  • Use - bullets. One change or reason per bullet.
  • Simple sentences. One subject per sentence. Max 2–3 short sentences per bullet — usually one is enough.
  • Say why when it isn't obvious from the diff, as its own plain sentence. No "because" or "so that" clauses, no parenthetical asides. Skip restating the obvious.
  • Describe the state the commit leaves. No history words: "now", "previously", "used to", "old code", "no longer".
  • Omit the body entirely for trivial commits — the subject is enough.

Breaking changes

  • Add ! after the type/scope: feat(api)!: drop v1 token format.
  • And/or a footer: BREAKING CHANGE: <what broke and the migration>.

One intent per commit

  • Default to one logical thing per commit. Don't mix a feature with an unrelated refactor.
  • If the message needs multiple feat:/fix: ideas, split into multiple commits.
  • A formatting sweep goes in its own style: commit, separate from behavior changes.
  • Exception: if the user explicitly asks to combine two changes into one commit, do it — keep the bulleted format and give each intent its own bullet (or small group) so the result is still easy to scan. Pick the subject type that covers the primary change.

Focus on what and why, not the test play-by-play

  • Describe what changed and why it matters — the behavior, the bug, the reason.
  • Don't narrate the testing regime in gory detail. "Add tests" or test: is enough; the diff shows the rest.
  • Mention tests only when it's the point of the commit (a test: commit) or when a test reveals something worth knowing (e.g. "covers the expired-token edge case that regressed").

Examples

Good:

fix(auth): reject expired refresh tokens

- Return 401 when a refresh token is past its TTL.
- Measure expiry from the TTL, not the issue time. A stale token must not refresh.
feat(cache): add per-entry expiry
refactor(parser): extract token reader into its own type

- Splits the 200-line scan loop so each token kind is testable.

Avoid:

  • A single commit titled update stuff mixing a feature, a fix, and reformatting.
  • A 6-paragraph body re-explaining the whole diff line by line.
  • Past tense / sentence-case subjects (Added the thing.).
  • A body bullet that runs four sentences and changes subject mid-bullet.

Checklist

  • Subject is type(scope): imperative subject, lowercase, no period
  • One intent per commit — unless the user asked to combine, then one bullet per intent
  • Body (if any) is - bullets, simple sentences, ≤ 2–3 per bullet
  • Body explains what and why, not the test play-by-play
  • Breaking changes marked with ! and/or BREAKING CHANGE: footer

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.

Reviews can't be read right now.

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.