Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Expert in building products that wrap AI APIs (OpenAI, Anthropic, etc. ) into focused tools people will pay for. Not just "ChatGPT but different" - products that solve specific problems with AI.
.claude/skills/sickn33-ai-wrapper-product/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | 151% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 147% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 178% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 218% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | 122% | 0% |
Expert in building products that wrap AI APIs (OpenAI, Anthropic, etc.) into focused tools people will pay for. Not just "ChatGPT but different" - products that solve specific problems with AI. Covers prompt engineering for products, cost management, rate limiting, and building defensible AI businesses.
Role: AI Product Architect
You know AI wrappers get a bad rap, but the good ones solve real problems. You build products where AI is the engine, not the gimmick. You understand prompt engineering is product development. You balance costs with user experience. You create AI products people actually pay for and use daily.
Read the detailed guide before executing this skill. It retains the complete procedure and reference material. Treat its safety, prerequisites, and validation requirements as mandatory. For focused work, load the relevant sections; for end-to-end work, read the guide completely.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 15,982 | 21,546 | +35% | 1 | 1 | 0% | 3,194 | 7,085 | +122% | 0 | 0 | — |
case-02 | fail→fail | 13,397 | 11,261 | -16% | 1 | 1 | 0% | 2,669 | 6,282 | +135% | 0 | 0 | — |
case-03 | pass→pass | 8,747 | 10,246 | +17% | 1 | 1 | 0% | 1,818 | 6,186 | +240% | 0 | 0 | — |
case-04 | fail→pass | 15,653 | 15,274 | -2% | 1 | 1 | 0% | 2,779 | 6,970 | +151% | 0 | 0 | — |
case-05 | pass→pass | 15,506 | 14,253 | -8% | 1 | 1 | 0% | 2,868 | 6,814 | +138% | 0 | 0 | — |
case-06 | pass→pass | 13,807 | 12,681 | -8% | 1 | 1 | 0% | 2,753 | 6,711 | +144% | 0 | 0 | — |
case-07 | fail→fail | 17,618 | 16,628 | -6% | 1 | 1 | 0% | 3,311 | 7,164 | +116% | 0 | 0 | — |
case-08 | pass→pass | 12,552 | 10,206 | -19% | 1 | 1 | 0% | 2,326 | 6,091 | +162% | 0 | 0 | — |
case-09 | fail→pass | 18,245 | 14,727 | -19% | 1 | 1 | 0% | 2,891 | 7,134 | +147% | 0 | 0 | — |
case-10 | pass→pass | 15,230 | 14,856 | -2% | 1 | 1 | 0% | 2,956 | 6,991 | +137% | 0 | 0 | — |
case-11 | fail→pass | 17,553 | 18,827 | +7% | 1 | 1 | 0% | 2,618 | 7,290 | +178% | 0 | 0 | — |
case-12 | pass→pass | 18,002 | 12,934 | -28% | 1 | 1 | 0% | 2,916 | 6,156 | +111% | 0 | 0 | — |
case-13 | fail→pass | 11,261 | 14,012 | +24% | 1 | 1 | 0% | 2,124 | 6,747 | +218% | 0 | 0 | — |
case-14 | fail→fail | 11,294 | 10,124 | -10% | 1 | 1 | 0% | 1,743 | 5,758 | +230% | 0 | 0 | — |
case-15 | pass→pass | 14,679 | 11,524 | -21% | 1 | 1 | 0% | 2,450 | 6,081 | +148% | 0 | 0 | — |
case-16 | pass→pass | 23,336 | 14,400 | -38% | 1 | 1 | 0% | 4,481 | 6,779 | +51% | 0 | 0 | — |
case-17 | pass→pass | 18,495 | 21,038 | +14% | 1 | 1 | 0% | 2,857 | 7,699 | +169% | 0 | 0 | — |
case-18 | pass→pass | 16,929 | 13,155 | -22% | 1 | 1 | 0% | 2,981 | 6,454 | +117% | 0 | 0 | — |
case-19 | pass→pass | 11,108 | 13,268 | +19% | 1 | 1 | 0% | 1,847 | 6,397 | +246% | 0 | 0 | — |
case-20 | pass→pass | 19,980 | 18,174 | -9% | 1 | 1 | 0% | 3,616 | 7,469 | +107% | 0 | 0 | — |
case-21 | pass→pass | 13,784 | 13,304 | -3% | 1 | 1 | 0% | 2,652 | 6,606 | +149% | 0 | 0 | — |
case-22 | pass→pass | 15,783 | 17,073 | +8% | 1 | 1 | 0% | 3,065 | 7,356 | +140% | 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 +18 percentage points is the difference between those two pass rates over the 22 comparable cases.
The publisher has shipped newer versions since this run, so these numbers describe v1, not the version currently listed.
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.