Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Build a throwaway prototype to answer a design question. Use when the user wants to sanity-check whether a state model or logic feels right, or explore what a UI should look like.
.claude/skills/fradser-prototype/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -36% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -30% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 8% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -27% | 0% |
| case-17 | ✗→✓ | ▲ Improved | -3% | 0% |
A prototype is throwaway code that answers a question. The question decides the shape.
Identify which question is being answered — from the user's prompt, the surrounding code, or by using the AskUserQuestion tool to ask the user if they're around:
The two branches produce very different artifacts — getting this wrong wastes the whole prototype. If the question is genuinely ambiguous and the user isn't reachable, default to whichever branch better matches the surrounding code (a backend module → logic; a page or component → UI) and state the assumption at the top of the prototype.
Identify the question being answered before building — "does this logic / state model feel right?" takes LOGIC.md, "what should this look like?" takes UI.md; getting this wrong wastes the whole prototype. Throwaway from day one: no tests, no polish, no persistence — keep the answer, delete the code.
pnpm <name>, python <path>, bun <path>, etc. The user must be able to start it without thinking.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-21 | pass→pass | 29,529 | 29,155 | -1% | 1 | 1 | 0% | 4,759 | 6,140 | +29% | 0 | 0 | — |
case-01 | fail→pass | 35,894 | 30,161 | -16% | 1 | 1 | 0% | 8,244 | 5,264 | -36% | 0 | 0 | — |
case-02 | fail→fail | 34,072 | 4,916 | -86% | 1 | 1 | 0% | 7,656 | 921 | -88% | 0 | 0 | — |
case-03 | fail→pass | 35,946 | 29,009 | -19% | 1 | 1 | 0% | 8,230 | 5,793 | -30% | 0 | 0 | — |
case-04 | pass→fail | 27,363 | 23,344 | -15% | 1 | 1 | 0% | 6,337 | 5,090 | -20% | 0 | 0 | — |
case-05 | fail→fail | 35,785 | 39,398 | +10% | 1 | 1 | 0% | 8,220 | 8,879 | +8% | 0 | 0 | — |
case-06 | fail→fail | 40,233 | 19,490 | -52% | 1 | 1 | 0% | 8,214 | 4,953 | -40% | 0 | 0 | — |
case-07 | pass→fail | 13,735 | 20,818 | +52% | 1 | 1 | 0% | 2,949 | 4,512 | +53% | 0 | 0 | — |
case-08 | fail→pass | 32,179 | 32,068 | -0% | 1 | 1 | 0% | 8,219 | 8,878 | +8% | 0 | 0 | — |
case-09 | pass→pass | 25,896 | 32,183 | +24% | 1 | 1 | 0% | 4,963 | 6,149 | +24% | 0 | 0 | — |
case-10 | fail→fail | 35,866 | 7,800 | -78% | 1 | 1 | 0% | 8,217 | 960 | -88% | 0 | 0 | — |
case-11 | fail→pass | 32,073 | 22,632 | -29% | 1 | 1 | 0% | 7,881 | 5,741 | -27% | 0 | 0 | — |
case-12 | fail→fail | 26,486 | 4,687 | -82% | 1 | 1 | 0% | 6,651 | 928 | -86% | 0 | 0 | — |
case-13 | fail→fail | 34,524 | 28,992 | -16% | 1 | 1 | 0% | 8,216 | 6,129 | -25% | 0 | 0 | — |
case-14 | fail→fail | 18,994 | 5,045 | -73% | 1 | 1 | 0% | 4,432 | 902 | -80% | 0 | 0 | — |
case-15 | fail→fail | 34,100 | 5,827 | -83% | 1 | 1 | 0% | 8,214 | 1,068 | -87% | 0 | 0 | — |
case-16 | fail→fail | 47,198 | 4,692 | -90% | 1 | 1 | 0% | 8,216 | 911 | -89% | 0 | 0 | — |
case-17 | fail→pass | 51,964 | 37,224 | -28% | 1 | 1 | 0% | 8,216 | 7,999 | -3% | 0 | 0 | — |
case-18 | fail→fail | 35,743 | 3,369 | -91% | 1 | 1 | 0% | 7,764 | 853 | -89% | 0 | 0 | — |
case-19 | pass→pass | 27,742 | 36,119 | +30% | 1 | 1 | 0% | 4,444 | 7,694 | +73% | 0 | 0 | — |
case-20 | fail→fail | 31,068 | 40,259 | +30% | 1 | 1 | 0% | 7,431 | 8,881 | +20% | 0 | 0 | — |
case-22 | fail→fail | 22,518 | 5,026 | -78% | 1 | 1 | 0% | 4,092 | 881 | -78% | 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 +14 percentage points is the difference between those two pass rates over the 14 comparable cases. 5 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.