git-commit

Generate commit messages. Adapt to repo style.

Safety Notice

This listing is imported from skills.sh public index metadata. Review upstream SKILL.md and repository scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "git-commit" with this command: npx skills add molechowski/claude-skills/molechowski-claude-skills-git-commit

Commit

Generate commit messages. Adapt to repo style.

Format

type(scope): description

[optional body]

[optional footer]

Workflow

  • Run git diff --staged to see changes

  • Run git log --oneline -10 to check repo style

  • Detect language from recent commits

  • Detect footer patterns (Change-Id, tickets)

  • Pick commit type from changes

  • Infer scope from affected files

  • Write short description

Language Detection

Check recent commits:

  • Polish: Dodanie, Naprawa, Poprawa, Zmiana, Usunięcie

  • English: Add, Fix, Update, Remove, Change

Match dominant language in last 10 commits.

Atomic Commits

One commit = one logical change.

Trigger rule: If changes need different types (feat + fix), split into separate commits.

Ask: Can this be described in one subject line without "and"?

Bad:

feat(ios): Add mTLS support and fix cache tests

Good (split into two):

feat(ios): Add mTLS certificate injection fix(ios): Fix cache tests for lowercase headers

Style Rules

  • Imperative mood: "add" not "added"

  • No period at end

  • No em dashes in prose

  • Under 50 chars for subject

  • Match repo language and conventions

Em dash vs hyphen:

  • Bad: "add auth - with token support"

  • Good: "add auth with token support"

Writing Style

Be laconic:

  • Shortest possible subject

  • No filler words

  • No bullet points in body

  • No verbose explanations

  • Skip body unless essential

Body is noise unless it adds value.

Subject examples:

  • Bad: "Add Error convenience extensions for NetworkError"

  • Good: "Add Error NetworkError extensions"

Body examples:

Bad (bullets):

feat(ios): Add mTLS support

  • Add certificate fields
  • Extract PEM data

Good (subject only):

feat(ios): Add mTLS certificate injection

Good (prose body when needed):

refactor(auth): Extract token service

Separate token logic for unit testing.

Body

Prefer no body. A good subject makes body unnecessary.

Decision tree:

  • Subject self-explanatory? Skip body

  • Multiple unrelated changes? Split commits

  • Need to explain "why"? Add prose body (1-2 sentences)

If body needed:

  • Prose only, no bullets

  • Explain why, not what

  • One sentence per logical change

  • Keep short

Breaking Changes

Add exclamation mark after type or scope:

feat!: remove deprecated API feat(api)!: change response format

Or use footer:

feat: update auth flow

BREAKING CHANGE: token format changed

Footers

Detect from repo history:

  • Change-Id (Gerrit): preserve if present

  • Ticket IDs (JIRA): include if pattern found

  • Signed-off-by: include if repo uses it

Do not add footers the repo does not use.

Safety

  • Check git status first

  • Never commit secrets

  • Verify staged files before commit

See quick-reference.md for type details.

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

General

res-price-compare

No summary provided by upstream source.

Repository SourceNeeds Review
General

res-deep

No summary provided by upstream source.

Repository SourceNeeds Review
General

doc-daily-digest

No summary provided by upstream source.

Repository SourceNeeds Review
General

doc-vault-project

No summary provided by upstream source.

Repository SourceNeeds Review