Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Helps write Git commit messages following the Conventional Commits specification. Use this skill when the user asks to commit changes, write commit messages, format commits, or mentions git commits.
.claude/skills/commit-helper/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | 52% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 102% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 34% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 127% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 267% | 0% |
A skill for creating properly formatted Git commit messages following the Conventional Commits specification.
This skill activates when you:
<type>(<scope>): <subject>
<body>
<footer>| Type | Description | |------|-------------| | feat | A new feature | | fix | A bug fix | | docs | Documentation only changes | | style | Changes that do not affect the meaning of the code (formatting, etc.) | | refactor | A code change that neither fixes a bug nor adds a feature | | perf | A code change that improves performance | | test | Adding missing tests or correcting existing tests | | chore | Changes to the build process or auxiliary tools | | ci | Changes to CI configuration files and scripts | | build | Changes that affect the build system or external dependencies |
The scope should indicate the area of the codebase affected:
components, hooks, store, styles, utilsapi, models, services, database, authci, deploy, dockerCloses #123, Fixes #456, Refs #789Closes #123, #456, #789BREAKING CHANGE: followed by descriptionfeat(auth): add OAuth2 login support
Implement OAuth2 authentication flow to allow users to log in
with their Google or GitHub accounts.
This change adds:
- New OAuth2 middleware for handling callbacks
- Updated login UI with social login buttons
- User profile synchronization
Closes #123fix(api): resolve race condition in user creation
The concurrent user creation requests could result in duplicate
email entries. Added unique constraint and proper error handling.
Fixes #456refactor(user): simplify profile update logic
Extracted common validation logic into a reusable function
to reduce code duplication across profile update endpoints.docs: update API documentation with new endpoints
Added documentation for the v2 user management endpoints
including request/response examples and error codes.updated stuff # Too vague, no type/scope
fixed bug # No context about which bug
feat: added feature # Redundant ("feat" means new feature)
Feat(User): Add Login # Incorrect capitalization
feat: A really really really long subject line that exceeds the recommended limit # Too longWhen introducing breaking changes, add BREAKING CHANGE: to the footer:
feat(api): migrate to REST v2
The API endpoints have been restructured for better consistency.
Old endpoints are deprecated and will be removed in v3.0.
BREAKING CHANGE: `/api/v1/users` is now `/api/v2/users`.
All consumers must update their integration by 2025-03-01.When writing a commit message:
git diff to understand what changedUse the validation script to check commit message format:
bashpython3 scripts/validate_commit.py "feat(scope): add behavior"
references/conventional-commits.md for full specificationreferences/examples.md for more examplesreferences/scopes.md for recommended scope naming| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 13,745 | 10,597 | -23% | 1 | 1 | 0% | 1,317 | 1,989 | +51% | 0 | 0 | — |
case-02 | fail→fail | 13,687 | 12,279 | -10% | 1 | 1 | 0% | 1,318 | 2,003 | +52% | 0 | 0 | — |
case-03 | fail→pass | 13,208 | 15,359 | +16% | 1 | 1 | 0% | 1,323 | 2,010 | +52% | 0 | 0 | — |
case-04 | pass→pass | 12,941 | 10,081 | -22% | 1 | 1 | 0% | 1,232 | 1,952 | +58% | 0 | 0 | — |
case-05 | pass→pass | 12,938 | 8,668 | -33% | 1 | 1 | 0% | 993 | 1,635 | +65% | 0 | 0 | — |
case-06 | fail→pass | 10,536 | 10,128 | -4% | 1 | 1 | 0% | 961 | 1,940 | +102% | 0 | 0 | — |
case-07 | fail→pass | 15,682 | 12,734 | -19% | 1 | 1 | 0% | 1,808 | 2,424 | +34% | 0 | 0 | — |
case-08 | fail→pass | 10,715 | 5,344 | -50% | 1 | 1 | 0% | 889 | 2,019 | +127% | 0 | 0 | — |
case-09 | fail→fail | 11,750 | 10,810 | -8% | 1 | 1 | 0% | 1,145 | 1,870 | +63% | 0 | 0 | — |
case-10 | pass→pass | 11,284 | 19,154 | +70% | 1 | 1 | 0% | 871 | 1,911 | +119% | 0 | 0 | — |
case-11 | fail→fail | 8,058 | 5,906 | -27% | 1 | 1 | 0% | 1,141 | 2,057 | +80% | 0 | 0 | — |
case-12 | pass→pass | 13,053 | 5,251 | -60% | 1 | 1 | 0% | 1,222 | 1,996 | +63% | 0 | 0 | — |
case-13 | fail→pass | 8,894 | 8,315 | -7% | 1 | 1 | 0% | 655 | 2,403 | +267% | 0 | 0 | — |
case-14 | fail→fail | 9,823 | 3,637 | -63% | 1 | 1 | 0% | 753 | 1,544 | +105% | 0 | 0 | — |
case-15 | fail→pass | 9,986 | 6,213 | -38% | 1 | 1 | 0% | 1,310 | 2,003 | +53% | 0 | 0 | — |
case-16 | fail→pass | 12,341 | 1,687 | -86% | 1 | 1 | 0% | 1,211 | 1,312 | +8% | 0 | 0 | — |
case-17 | fail→fail | 16,048 | 16,748 | +4% | 1 | 1 | 0% | 1,438 | 2,098 | +46% | 0 | 0 | — |
case-18 | fail→fail | 4,861 | 9,201 | +89% | 1 | 1 | 0% | 787 | 1,768 | +125% | 0 | 0 | — |
case-19 | fail→fail | 10,507 | 6,194 | -41% | 1 | 1 | 0% | 963 | 2,228 | +131% | 0 | 0 | — |
case-20 | pass→fail | 15,661 | 5,489 | -65% | 1 | 1 | 0% | 1,024 | 2,024 | +98% | 0 | 0 | — |
case-21 | pass→pass | 11,795 | 8,370 | -29% | 1 | 1 | 0% | 1,995 | 2,654 | +33% | 0 | 0 | — |
case-22 | pass→pass | 11,802 | 5,607 | -52% | 1 | 1 | 0% | 1,179 | 1,897 | +61% | 0 | 0 | — |
case-23 | pass→pass | 8,318 | 12,029 | +45% | 1 | 1 | 0% | 1,454 | 2,068 | +42% | 0 | 0 | — |
case-24 | pass→fail | 6,353 | 9,384 | +48% | 1 | 1 | 0% | 1,064 | 1,834 | +72% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 24 cases were attempted. The headline lift of +21 percentage points is the difference between those two pass rates over the 24 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 7/24/2026 | +23% |
Other measured skills in the registry, with their headline benchmark lift.