Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Analyzes code and project structure to find optimization opportunities without making changes. Triggers when users say things like "optimize my code", "review my code for improvements", "find inefficiencies in my code", "suggest code improvements", "analyze my code quality", "check for complexity issues", "find unused code", "code review suggestions", or "how can I improve my code". Produces labeled suggestions: [Delete] for unused/dead code, [Improve] for better implementations, and [Complexity
.claude/skills/oykunehir-code-optimization-suggestions/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 46% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -16% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -1% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 7% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 70% | 0% |
Analyze the codebase and provide structured suggestions without modifying any files.
[Delete] `functionName` in <filename>, line <N>–<N>: This function is defined but never called anywhere in the project. Safe to remove.
[Improve] <filename>, line <N>: You currently have:
<existing code snippet>
Consider replacing with:
<improved code snippet>
Reason: <brief explanation>
[Complexity] <filename>, line <N>–<N>: This section runs at O(n²) due to nested loops over the same collection. Consider <alternative approach> to reduce it to O(n log n) or O(n).| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 5,208 | 6,297 | +21% | 1 | 1 | 0% | 906 | 639 | -29% | 0 | 0 | — |
case-02 | fail→pass | 6,770 | 7,110 | +5% | 1 | 1 | 0% | 1,153 | 1,685 | +46% | 0 | 0 | — |
case-03 | fail→pass | 7,724 | 4,760 | -38% | 1 | 1 | 0% | 1,520 | 1,279 | -16% | 0 | 0 | — |
case-04 | fail→pass | 10,396 | 7,811 | -25% | 1 | 1 | 0% | 1,964 | 1,953 | -1% | 0 | 0 | — |
case-05 | fail→fail | 2,961 | 9,347 | +216% | 1 | 1 | 0% | 92 | 1,159 | +1160% | 0 | 0 | — |
case-06 | fail→fail | 8,757 | 4,354 | -50% | 1 | 1 | 0% | 1,578 | 717 | -55% | 0 | 0 | — |
case-07 | fail→fail | 12,165 | 3,095 | -75% | 1 | 1 | 0% | 2,224 | 590 | -73% | 0 | 0 | — |
case-08 | pass→pass | 16,366 | 13,255 | -19% | 1 | 1 | 0% | 2,921 | 3,061 | +5% | 0 | 0 | — |
case-09 | pass→pass | 7,322 | 3,348 | -54% | 1 | 1 | 0% | 1,262 | 1,099 | -13% | 0 | 0 | — |
case-10 | fail→fail | 2,766 | 1,913 | -31% | 1 | 1 | 0% | 463 | 648 | +40% | 0 | 0 | — |
case-11 | pass→fail | 16,189 | 2,216 | -86% | 1 | 1 | 0% | 3,203 | 560 | -83% | 0 | 0 | — |
case-12 | pass→fail | 11,901 | 3,480 | -71% | 1 | 1 | 0% | 2,170 | 773 | -64% | 0 | 0 | — |
case-13 | fail→fail | 6,292 | 3,651 | -42% | 1 | 1 | 0% | 954 | 540 | -43% | 0 | 0 | — |
case-14 | fail→pass | 6,945 | 10,853 | +56% | 1 | 1 | 0% | 1,312 | 1,401 | +7% | 0 | 0 | — |
case-15 | fail→fail | 14,538 | 3,967 | -73% | 1 | 1 | 0% | 2,130 | 624 | -71% | 0 | 0 | — |
case-16 | fail→pass | 4,372 | 3,268 | -25% | 1 | 1 | 0% | 625 | 1,062 | +70% | 0 | 0 | — |
case-17 | fail→pass | 5,214 | 7,020 | +35% | 1 | 1 | 0% | 931 | 1,117 | +20% | 0 | 0 | — |
case-18 | pass→pass | 4,516 | 4,162 | -8% | 1 | 1 | 0% | 602 | 1,108 | +84% | 0 | 0 | — |
case-19 | pass→pass | 8,448 | 8,598 | +2% | 1 | 1 | 0% | 1,489 | 1,859 | +25% | 0 | 0 | — |
case-20 | pass→fail | 19,793 | 2,458 | -88% | 1 | 1 | 0% | 4,919 | 540 | -89% | 0 | 0 | — |
case-21 | fail→fail | 4,355 | 1,869 | -57% | 1 | 1 | 0% | 740 | 665 | -10% | 0 | 0 | — |
case-22 | pass→fail | 5,112 | 2,215 | -57% | 1 | 1 | 0% | 825 | 721 | -13% | 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 14 counted toward the lift figure. The other 8 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 +9 percentage points is the difference between those two pass rates over the 14 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.