Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Run the repo's full validation suite (format, build, size, tests, e2e). Invoke after any important change — finishing a feature, fix, refactor, or dependency bump — and always before committing.
.claude/skills/itsbencodes-validate/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | -53% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -46% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -63% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -67% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -2% | 0% |
Run every step.
bashnpm run format npm run build npm run size npm test npx playwright install npx playwright test
If a step fails, fix the cause and re-run the full suite from the top — earlier steps may now produce different output (formatting, generated files, etc.). Repeat until a clean pass with no intervening changes.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-06 | fail→fail | 13,238 | 10,375 | -22% | 1 | 1 | 0% | 1,958 | 616 | -69% | 0 | 0 | — |
case-01 | fail→fail | 42,446 | 10,043 | -76% | 1 | 1 | 0% | 174 | 293 | +68% | 0 | 0 | — |
case-02 | fail→fail | 5,768 | 4,840 | -16% | 1 | 1 | 0% | 235 | 223 | -5% | 0 | 0 | — |
case-03 | fail→fail | 6,066 | 5,302 | -13% | 1 | 1 | 0% | 179 | 264 | +47% | 0 | 0 | — |
case-04 | fail→fail | 9,769 | 9,587 | -2% | 1 | 1 | 0% | 1,498 | 468 | -69% | 0 | 0 | — |
case-05 | fail→fail | 9,308 | 5,601 | -40% | 1 | 1 | 0% | 1,448 | 753 | -48% | 0 | 0 | — |
case-07 | fail→pass | 7,266 | 2,920 | -60% | 1 | 1 | 0% | 1,062 | 497 | -53% | 0 | 0 | — |
case-08 | fail→pass | 10,905 | 5,037 | -54% | 1 | 1 | 0% | 1,498 | 806 | -46% | 0 | 0 | — |
case-09 | pass→pass | 9,802 | 2,736 | -72% | 1 | 1 | 0% | 1,087 | 561 | -48% | 0 | 0 | — |
case-10 | fail→pass | 12,328 | 8,606 | -30% | 1 | 1 | 0% | 1,735 | 639 | -63% | 0 | 0 | — |
case-11 | fail→pass | 9,704 | 2,606 | -73% | 1 | 1 | 0% | 1,325 | 435 | -67% | 0 | 0 | — |
case-12 | pass→pass | 13,267 | 12,053 | -9% | 1 | 1 | 0% | 1,748 | 1,916 | +10% | 0 | 0 | — |
case-13 | fail→pass | 5,103 | 3,947 | -23% | 1 | 1 | 0% | 613 | 598 | -2% | 0 | 0 | — |
case-14 | fail→pass | 12,693 | 10,429 | -18% | 1 | 1 | 0% | 1,960 | 1,476 | -25% | 0 | 0 | — |
case-15 | fail→pass | 10,028 | 4,995 | -50% | 1 | 1 | 0% | 1,558 | 805 | -48% | 0 | 0 | — |
case-16 | pass→fail | 6,418 | 6,099 | -5% | 1 | 1 | 0% | 1,176 | 296 | -75% | 0 | 0 | — |
case-17 | pass→fail | 32,496 | 5,101 | -84% | 1 | 1 | 0% | 1,051 | 259 | -75% | 0 | 0 | — |
case-18 | pass→fail | 81,940 | 224,185 | +174% | 1 | 1 | 0% | 842 | 407 | -52% | 0 | 0 | — |
case-19 | pass→fail | 9,587 | 6,585 | -31% | 1 | 1 | 0% | 1,324 | 254 | -81% | 0 | 0 | — |
case-20 | fail→pass | 11,026 | 7,793 | -29% | 1 | 1 | 0% | 1,536 | 438 | -71% | 0 | 0 | — |
case-21 | fail→pass | 6,717 | 2,869 | -57% | 1 | 1 | 0% | 1,114 | 528 | -53% | 0 | 0 | — |
case-22 | pass→pass | 4,320 | 5,127 | +19% | 1 | 1 | 0% | 602 | 739 | +23% | 0 | 0 | — |
case-23 | fail→pass | 12,484 | 18,991 | +52% | 1 | 1 | 0% | 1,222 | 2,588 | +112% | 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. 23 cases were attempted, and 14 counted toward the lift figure. The other 9 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 +26 percentage points is the difference between those two pass rates over the 14 comparable cases. 4 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.