Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Conventional Commits 1.0 generator. Stages relevant files, infers type(scope): description, never uses --no-verify.
.claude/skills/hashgraph-online-commit/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | 18% | 0% |
| case-16 | ✗→✓ | ▲ Improved | -51% | 0% |
| case-04 | ✓→✗ | ▼ Worse | -47% | 0% |
| case-05 | ✓→✗ | ▼ Worse | -61% | 0% |
| case-06 | ✓→✗ | ▼ Worse | -5% | 0% |
Inspect git diff --cached and git log --oneline -5, write a Conventional Commits subject (body only for breaking changes or non-obvious reasoning), commit, verify.
git status — never use -uall.Co-Authored-By or any AI-attribution trailer (e.g. "Generated with Claude Code"). Overrides default Claude Code behavior.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-07 | fail→fail | 17,342 | 14,647 | -16% | 1 | 1 | 0% | 415 | 358 | -14% | 0 | 0 | — |
case-01 | fail→fail | 15,271 | 38,988 | +155% | 1 | 1 | 0% | 228 | 297 | +30% | 0 | 0 | — |
case-02 | fail→fail | 13,654 | 13,353 | -2% | 1 | 1 | 0% | 1,235 | 279 | -77% | 0 | 0 | — |
case-03 | fail→fail | 17,556 | 13,194 | -25% | 1 | 1 | 0% | 165 | 184 | +12% | 0 | 0 | — |
case-04 | pass→fail | 8,550 | 14,724 | +72% | 1 | 1 | 0% | 630 | 331 | -47% | 0 | 0 | — |
case-05 | pass→fail | 11,567 | 13,779 | +19% | 1 | 1 | 0% | 629 | 243 | -61% | 0 | 0 | — |
case-06 | pass→fail | 7,374 | 17,510 | +137% | 1 | 1 | 0% | 362 | 345 | -5% | 0 | 0 | — |
case-08 | fail→pass | 9,337 | 11,466 | +23% | 1 | 1 | 0% | 1,132 | 1,337 | +18% | 0 | 0 | — |
case-09 | fail→fail | 14,375 | 9,957 | -31% | 1 | 1 | 0% | 1,423 | 811 | -43% | 0 | 0 | — |
case-10 | pass→fail | 10,093 | 16,032 | +59% | 1 | 1 | 0% | 728 | 383 | -47% | 0 | 0 | — |
case-11 | fail→fail | 10,062 | 15,579 | +55% | 1 | 1 | 0% | 218 | 492 | +126% | 0 | 0 | — |
case-17 | pass→pass | 10,713 | 8,415 | -21% | 1 | 1 | 0% | 888 | 1,393 | +57% | 0 | 0 | — |
case-12 | fail→fail | 10,263 | 9,051 | -12% | 1 | 1 | 0% | 764 | 370 | -52% | 0 | 0 | — |
case-13 | fail→fail | 12,879 | 15,616 | +21% | 1 | 1 | 0% | 376 | 322 | -14% | 0 | 0 | — |
case-14 | fail→fail | 9,153 | 3,983 | -56% | 1 | 1 | 0% | 177 | 339 | +92% | 0 | 0 | — |
case-15 | fail→fail | 19,329 | 48,831 | +153% | 1 | 1 | 0% | 584 | 313 | -46% | 0 | 0 | — |
case-16 | fail→pass | 8,004 | 4,289 | -46% | 1 | 1 | 0% | 1,287 | 636 | -51% | 0 | 0 | — |
case-18 | pass→fail | 9,638 | 9,255 | -4% | 1 | 1 | 0% | 1,509 | 410 | -73% | 0 | 0 | — |
case-19 | fail→fail | 7,714 | 9,541 | +24% | 1 | 1 | 0% | 301 | 298 | -1% | 0 | 0 | — |
case-20 | fail→fail | 2,936 | 19,746 | +573% | 1 | 1 | 0% | 456 | 286 | -37% | 0 | 0 | — |
case-21 | fail→fail | 9,700 | 10,056 | +4% | 1 | 1 | 0% | 651 | 237 | -64% | 0 | 0 | — |
case-22 | pass→fail | 5,270 | 4,515 | -14% | 1 | 1 | 0% | 900 | 284 | -68% | 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, and 4 counted toward the lift figure. The other 18 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 -18 percentage points is the difference between those two pass rates over the 4 comparable cases. 6 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.
Other measured skills in the registry, with their headline benchmark lift.