Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Analyze staged/unstaged changes and create semantic conventional commits with context about WHY, not just WHAT. Auto-detects commit type and scope from the diff. Supports optional type/scope arguments. Usage - /commit-smart, /commit-smart fix, /commit-smart refactor api
.claude/skills/davila7-commit-smart/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | 34% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 48% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 57% | 0% |
| case-23 | ✓→✗ | ▼ Worse | -42% | 0% |
| case-04 | ✓→✓ | = Same ✓ | 109% | 0% |
Create meaningful conventional commits by analyzing your actual changes.
Run these commands to understand the current state:
bashgit status git diff --stat git diff --cached --stat
If nothing is staged (git diff --cached is empty):
git add <files>If changes are already staged, proceed to analysis.
Read the full staged diff:
bashgit diff --cached
Determine the commit type from the changes:
| Signal | Type | |--------|------| | New files with new functionality | feat | | New test files or test additions | test | | Changes to existing logic fixing incorrect behavior | fix | | Structural changes without behavior change | refactor | | package.json, tsconfig, CI config changes | chore | | Build/bundler config changes | build | | README, docs, comments only | docs | | Formatting, whitespace, semicolons only | style | | Performance improvements | perf |
Determine the scope from the primary directory or module affected:
src/api/ -> apisrc/components/auth/ -> authtests/ -> testsIf the user provided arguments via $ARGUMENTS:
fix) -> use as commit typerefactor api) -> use as type and scopeFormat: type(scope): imperative short description
Rules:
Example:
feat(auth): add JWT refresh token rotation
Tokens were expiring mid-session for users with slow connections.
Rotating refresh tokens extends the session without compromising
security, since each refresh token can only be used once.Show the user the proposed commit message and ask for confirmation.
If confirmed, run:
bashgit commit -m "<message>"
Then verify with:
bashgit log --oneline -1
Show the committed hash and message.
! after the scope: feat(api)!: change response format| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 2,881 | 2,353 | -18% | 1 | 1 | 0% | 470 | 925 | +97% | 0 | 0 | — |
case-02 | fail→fail | 2,504 | 3,311 | +32% | 1 | 1 | 0% | 291 | 906 | +211% | 0 | 0 | — |
case-03 | fail→fail | 2,056 | 1,720 | -16% | 1 | 1 | 0% | 310 | 906 | +192% | 0 | 0 | — |
case-04 | pass→pass | 3,224 | 2,355 | -27% | 1 | 1 | 0% | 537 | 1,125 | +109% | 0 | 0 | — |
case-05 | fail→pass | 5,756 | 3,762 | -35% | 1 | 1 | 0% | 1,018 | 1,365 | +34% | 0 | 0 | — |
case-06 | pass→pass | 8,819 | 3,262 | -63% | 1 | 1 | 0% | 1,416 | 1,341 | -5% | 0 | 0 | — |
case-07 | fail→pass | 4,429 | 3,032 | -32% | 1 | 1 | 0% | 834 | 1,234 | +48% | 0 | 0 | — |
case-08 | pass→pass | 6,892 | 5,101 | -26% | 1 | 1 | 0% | 1,139 | 1,583 | +39% | 0 | 0 | — |
case-09 | pass→pass | 9,162 | 5,828 | -36% | 1 | 1 | 0% | 1,607 | 1,759 | +9% | 0 | 0 | — |
case-10 | pass→pass | 7,884 | 3,879 | -51% | 1 | 1 | 0% | 1,253 | 1,448 | +16% | 0 | 0 | — |
case-11 | pass→pass | 5,270 | 4,070 | -23% | 1 | 1 | 0% | 992 | 1,456 | +47% | 0 | 0 | — |
case-12 | pass→pass | 6,477 | 3,118 | -52% | 1 | 1 | 0% | 972 | 1,251 | +29% | 0 | 0 | — |
case-13 | fail→fail | 9,259 | 5,205 | -44% | 1 | 1 | 0% | 1,628 | 1,565 | -4% | 0 | 0 | — |
case-14 | pass→pass | 11,238 | 4,588 | -59% | 1 | 1 | 0% | 1,769 | 1,525 | -14% | 0 | 0 | — |
case-15 | fail→pass | 4,200 | 1,832 | -56% | 1 | 1 | 0% | 656 | 1,028 | +57% | 0 | 0 | — |
case-16 | fail→fail | 5,052 | 4,237 | -16% | 1 | 1 | 0% | 758 | 1,465 | +93% | 0 | 0 | — |
case-17 | pass→pass | 5,185 | 2,588 | -50% | 1 | 1 | 0% | 942 | 1,118 | +19% | 0 | 0 | — |
case-18 | pass→pass | 4,089 | 2,513 | -39% | 1 | 1 | 0% | 895 | 1,186 | +33% | 0 | 0 | — |
case-19 | pass→pass | 5,151 | 2,621 | -49% | 1 | 1 | 0% | 881 | 1,151 | +31% | 0 | 0 | — |
case-20 | pass→pass | 2,962 | 2,458 | -17% | 1 | 1 | 0% | 508 | 1,056 | +108% | 0 | 0 | — |
case-21 | pass→pass | 9,760 | 5,848 | -40% | 1 | 1 | 0% | 1,800 | 1,908 | +6% | 0 | 0 | — |
case-22 | fail→fail | 2,801 | 1,179 | -58% | 1 | 1 | 0% | 532 | 911 | +71% | 0 | 0 | — |
case-23 | pass→fail | 8,062 | 4,578 | -43% | 1 | 1 | 0% | 1,637 | 946 | -42% | 0 | 0 | — |
case-24 | pass→pass | 10,819 | 5,449 | -50% | 1 | 1 | 0% | 2,196 | 1,831 | -17% | 0 | 0 | — |
case-25 | pass→pass | 8,000 | 8,514 | +6% | 1 | 1 | 0% | 1,409 | 1,932 | +37% | 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. 25 cases were attempted, and 24 counted toward the lift figure. The other 1 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +8 percentage points is the difference between those two pass rates over the 24 comparable cases. 1 case got worse with the skill loaded, and it is 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.
Other measured skills in the registry, with their headline benchmark lift.