Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →General principles for exploring, developing, editing, and refactoring code. Use for codebase analysis, implementation tasks, and code quality improvements.
.claude/skills/aiskillstore-develop/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | -10% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -78% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -66% | 0% |
| case-22 | ✗→✓ | ▲ Improved | 45% | 0% |
| case-08 | ✓→✗ | ▼ Worse | -11% | 0% |
For non-trivial changes:
plan skill.When working with specific languages, consult the following references:
references/python.md.edit_file: Use for targeted changes; avoid rewriting full files.fd and rg for exploration; standard shell tools for navigation.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 13,097 | 5,297 | -60% | 1 | 1 | 0% | 2,113 | 554 | -74% | 0 | 0 | — |
case-07 | fail→pass | 9,710 | 7,068 | -27% | 1 | 1 | 0% | 1,792 | 1,608 | -10% | 0 | 0 | — |
case-02 | fail→fail | 3,855 | 21,120 | +448% | 1 | 1 | 0% | 484 | 554 | +14% | 0 | 0 | — |
case-03 | fail→fail | 10,397 | 4,980 | -52% | 1 | 1 | 0% | 1,664 | 841 | -49% | 0 | 0 | — |
case-04 | fail→fail | 4,294 | 31,283 | +629% | 1 | 1 | 0% | 343 | 5,637 | +1543% | 0 | 0 | — |
case-05 | fail→fail | 6,676 | 7,794 | +17% | 1 | 1 | 0% | 784 | 613 | -22% | 0 | 0 | — |
case-06 | fail→pass | 23,207 | 1,748 | -92% | 1 | 1 | 0% | 2,226 | 495 | -78% | 0 | 0 | — |
case-08 | pass→fail | 4,129 | 2,201 | -47% | 1 | 1 | 0% | 652 | 578 | -11% | 0 | 0 | — |
case-09 | pass→pass | 5,561 | 4,000 | -28% | 1 | 1 | 0% | 845 | 637 | -25% | 0 | 0 | — |
case-10 | pass→pass | 13,036 | 2,848 | -78% | 1 | 1 | 0% | 2,259 | 683 | -70% | 0 | 0 | — |
case-11 | pass→fail | 14,788 | 26,393 | +78% | 1 | 1 | 0% | 2,290 | 435 | -81% | 0 | 0 | — |
case-12 | pass→pass | 5,030 | 3,253 | -35% | 1 | 1 | 0% | 733 | 716 | -2% | 0 | 0 | — |
case-13 | fail→pass | 10,729 | 2,309 | -78% | 1 | 1 | 0% | 1,668 | 570 | -66% | 0 | 0 | — |
case-14 | pass→pass | 14,407 | 8,197 | -43% | 1 | 1 | 0% | 1,659 | 1,565 | -6% | 0 | 0 | — |
case-15 | pass→pass | 9,447 | 5,517 | -42% | 1 | 1 | 0% | 1,414 | 1,104 | -22% | 0 | 0 | — |
case-16 | pass→fail | 4,664 | 5,785 | +24% | 1 | 1 | 0% | 668 | 429 | -36% | 0 | 0 | — |
case-17 | fail→fail | 1,614 | 7,278 | +351% | 1 | 1 | 0% | 219 | 653 | +198% | 0 | 0 | — |
case-18 | pass→pass | 12,426 | 6,604 | -47% | 1 | 1 | 0% | 1,811 | 687 | -62% | 0 | 0 | — |
case-19 | pass→pass | 13,729 | 6,682 | -51% | 1 | 1 | 0% | 2,249 | 1,447 | -36% | 0 | 0 | — |
case-20 | pass→pass | 15,022 | 14,110 | -6% | 1 | 1 | 0% | 2,760 | 2,526 | -8% | 0 | 0 | — |
case-21 | pass→pass | 11,117 | 8,141 | -27% | 1 | 1 | 0% | 1,846 | 1,631 | -12% | 0 | 0 | — |
case-22 | fail→pass | 2,246 | 2,440 | +9% | 1 | 1 | 0% | 378 | 549 | +45% | 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 16 counted toward the lift figure. The other 6 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 +5 percentage points is the difference between those two pass rates over the 16 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.