Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Compare current benchmark results against historical baselines to spot performance regressions. Use as an explicit/manual helper for build-to-build degradation review, not for broad optimization strategy or low-level profiling ownership.
.claude/skills/foryourhealth111-pixel-detecting-performance-regressions/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-15 | ✗→✓ | ▲ Improved | -16% | 0% |
| case-03 | ✓→✗ | ▼ Worse | -28% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | -14% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | 67% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 24% | 0% |
Treat this skill as an explicit/manual helper for benchmark comparison work.
Use this skill when:
providing-performance-optimization-advicesystematic-debugginggenerating-test-reportsperformance-testing for benchmark generationproviding-performance-optimization-advice after the regression is confirmed| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 10,387 | 8,410 | -19% | 1 | 1 | 0% | 1,919 | 1,650 | -14% | 0 | 0 | — |
case-02 | pass→pass | 6,965 | 11,680 | +68% | 1 | 1 | 0% | 1,424 | 2,384 | +67% | 0 | 0 | — |
case-03 | pass→fail | 12,922 | 7,945 | -39% | 1 | 1 | 0% | 2,150 | 1,554 | -28% | 0 | 0 | — |
case-04 | pass→pass | 5,494 | 5,359 | -2% | 1 | 1 | 0% | 906 | 1,124 | +24% | 0 | 0 | — |
case-05 | pass→pass | 12,152 | 11,696 | -4% | 1 | 1 | 0% | 2,235 | 1,989 | -11% | 0 | 0 | — |
case-06 | pass→pass | 14,359 | 9,108 | -37% | 1 | 1 | 0% | 2,831 | 1,774 | -37% | 0 | 0 | — |
case-07 | pass→pass | 15,029 | 9,957 | -34% | 1 | 1 | 0% | 2,745 | 2,038 | -26% | 0 | 0 | — |
case-08 | pass→pass | 11,918 | 11,893 | -0% | 1 | 1 | 0% | 2,180 | 2,297 | +5% | 0 | 0 | — |
case-09 | fail→fail | 19,762 | 11,600 | -41% | 1 | 1 | 0% | 3,191 | 2,169 | -32% | 0 | 0 | — |
case-10 | pass→pass | 15,659 | 10,257 | -34% | 1 | 1 | 0% | 2,602 | 1,906 | -27% | 0 | 0 | — |
case-11 | pass→pass | 12,874 | 10,667 | -17% | 1 | 1 | 0% | 2,141 | 1,891 | -12% | 0 | 0 | — |
case-12 | fail→fail | 11,689 | 7,336 | -37% | 1 | 1 | 0% | 2,011 | 1,369 | -32% | 0 | 0 | — |
case-13 | pass→pass | 15,118 | 7,863 | -48% | 1 | 1 | 0% | 2,305 | 1,448 | -37% | 0 | 0 | — |
case-14 | pass→pass | 11,468 | 8,289 | -28% | 1 | 1 | 0% | 2,035 | 1,593 | -22% | 0 | 0 | — |
case-15 | fail→pass | 12,953 | 9,746 | -25% | 1 | 1 | 0% | 2,107 | 1,776 | -16% | 0 | 0 | — |
case-16 | pass→pass | 15,869 | 9,697 | -39% | 1 | 1 | 0% | 2,683 | 1,821 | -32% | 0 | 0 | — |
case-17 | pass→pass | 15,082 | 7,838 | -48% | 1 | 1 | 0% | 2,364 | 1,524 | -36% | 0 | 0 | — |
case-18 | fail→fail | 19,390 | 11,070 | -43% | 1 | 1 | 0% | 3,083 | 2,062 | -33% | 0 | 0 | — |
case-19 | pass→pass | 13,637 | 8,478 | -38% | 1 | 1 | 0% | 2,183 | 1,532 | -30% | 0 | 0 | — |
case-20 | pass→pass | 18,484 | 9,260 | -50% | 1 | 1 | 0% | 3,402 | 1,853 | -46% | 0 | 0 | — |
case-21 | pass→pass | 16,652 | 9,920 | -40% | 1 | 1 | 0% | 3,165 | 2,160 | -32% | 0 | 0 | — |
case-22 | pass→pass | 18,808 | 11,458 | -39% | 1 | 1 | 0% | 4,592 | 2,620 | -43% | 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 0 percentage points is the difference between those two pass rates over the 22 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.