Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Generate well-formatted git commit messages following conventional commit standards
.claude/skills/thinkinaixyz-git-commit/SKILL.md| Model | Eval pass | Runs |
|---|---|---|
| gemini-3.6-flash | 100% | 8 |
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | -16% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -4% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 13% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 68% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | 56% | 0% |
You are a git commit message expert. When this skill is activated, help users create well-structured commit messages.
Follow the Conventional Commits specification:
<type>(<scope>): <subject>
[optional body]
[optional footer(s)]Fixes #123BREAKING CHANGE: descriptiongit diff --staged or git status to see changesgit commit -m "message" if user confirms| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-04 | pass→pass | 2,500 | 2,490 | -0% | 1 | 1 | 0% | 450 | 755 | +68% | 0 | 0 | — |
case-03 | fail→pass | 6,939 | 4,036 | -42% | 1 | 1 | 0% | 1,233 | 1,031 | -16% | 0 | 0 | — |
case-01 | pass→pass | 3,683 | 3,258 | -12% | 1 | 1 | 0% | 591 | 923 | +56% | 0 | 0 | — |
case-02 | fail→pass | 5,223 | 2,924 | -44% | 1 | 1 | 0% | 936 | 900 | -4% | 0 | 0 | — |
case-05 | pass→pass | 3,358 | 2,809 | -16% | 1 | 1 | 0% | 642 | 814 | +27% | 0 | 0 | — |
case-06 | pass→pass | 3,958 | 2,858 | -28% | 1 | 1 | 0% | 731 | 823 | +13% | 0 | 0 | — |
case-07 | pass→pass | 3,114 | 2,819 | -9% | 1 | 1 | 0% | 499 | 802 | +61% | 0 | 0 | — |
case-08 | pass→pass | 4,567 | 3,233 | -29% | 1 | 1 | 0% | 781 | 977 | +25% | 0 | 0 | — |
case-09 | pass→pass | 3,034 | 65,230 | +2050% | 1 | 1 | 0% | 509 | 1,051 | +106% | 0 | 0 | — |
case-10 | pass→pass | 5,895 | 3,625 | -39% | 1 | 1 | 0% | 882 | 968 | +10% | 0 | 0 | — |
case-11 | pass→pass | 4,863 | 3,278 | -33% | 1 | 1 | 0% | 858 | 964 | +12% | 0 | 0 | — |
case-12 | fail→pass | 4,495 | 3,349 | -25% | 1 | 1 | 0% | 729 | 823 | +13% | 0 | 0 | — |
case-13 | pass→pass | 3,439 | 3,726 | +8% | 1 | 1 | 0% | 615 | 980 | +59% | 0 | 0 | — |
case-14 | pass→pass | 3,550 | 3,552 | +0% | 1 | 1 | 0% | 582 | 991 | +70% | 0 | 0 | — |
case-15 | pass→pass | 6,090 | 5,050 | -17% | 1 | 1 | 0% | 984 | 1,157 | +18% | 0 | 0 | — |
case-16 | pass→pass | 7,982 | 6,454 | -19% | 1 | 1 | 0% | 1,347 | 1,357 | +1% | 0 | 0 | — |
case-17 | pass→pass | 2,491 | 1,843 | -26% | 1 | 1 | 0% | 342 | 671 | +96% | 0 | 0 | — |
case-18 | pass→pass | 2,579 | 1,965 | -24% | 1 | 1 | 0% | 435 | 667 | +53% | 0 | 0 | — |
case-19 | pass→pass | 2,637 | 3,462 | +31% | 1 | 1 | 0% | 457 | 935 | +105% | 0 | 0 | — |
case-20 | pass→pass | 13,305 | 9,290 | -30% | 1 | 1 | 0% | 2,290 | 1,972 | -14% | 0 | 0 | — |
case-21 | pass→pass | 5,218 | 3,561 | -32% | 1 | 1 | 0% | 907 | 959 | +6% | 0 | 0 | — |
case-22 | pass→pass | 5,630 | 3,141 | -44% | 1 | 1 | 0% | 985 | 920 | -7% | 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. 22 cases were attempted. The headline lift of +14 percentage points is the difference between those two pass rates over the 22 comparable cases.
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.