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/mxyhi-prototype/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 10% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 102% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -2% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -6% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -13% | 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 asking if the user is 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.
pnpm <name>, python <path>, bun <path>, etc. A logic demo is a single HTML file the user double-clicks. Either way, no thinking required to start it.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-06 | pass→pass | 17,135 | 12,255 | -28% | 1 | 1 | 0% | 2,802 | 2,330 | -17% | 0 | 0 | — |
case-01 | fail→fail | 25,601 | 25,341 | -1% | 1 | 1 | 0% | 6,195 | 6,807 | +10% | 0 | 0 | — |
case-02 | fail→pass | 25,855 | 26,129 | +1% | 1 | 1 | 0% | 6,182 | 6,794 | +10% | 0 | 0 | — |
case-03 | fail→pass | 15,258 | 26,202 | +72% | 1 | 1 | 0% | 3,371 | 6,797 | +102% | 0 | 0 | — |
case-04 | pass→pass | 11,836 | 11,158 | -6% | 1 | 1 | 0% | 2,423 | 2,821 | +16% | 0 | 0 | — |
case-05 | pass→pass | 18,825 | 14,420 | -23% | 1 | 1 | 0% | 4,069 | 3,806 | -6% | 0 | 0 | — |
case-07 | fail→fail | 19,916 | 28,012 | +41% | 1 | 1 | 0% | 4,453 | 6,789 | +52% | 0 | 0 | — |
case-08 | fail→pass | 26,276 | 24,497 | -7% | 1 | 1 | 0% | 6,179 | 6,051 | -2% | 0 | 0 | — |
case-09 | pass→pass | 9,180 | 6,525 | -29% | 1 | 1 | 0% | 1,798 | 1,895 | +5% | 0 | 0 | — |
case-10 | fail→pass | 11,812 | 8,159 | -31% | 1 | 1 | 0% | 2,063 | 1,944 | -6% | 0 | 0 | — |
case-11 | fail→pass | 10,678 | 5,584 | -48% | 1 | 1 | 0% | 1,657 | 1,449 | -13% | 0 | 0 | — |
case-12 | pass→pass | 10,310 | 5,096 | -51% | 1 | 1 | 0% | 1,580 | 1,367 | -13% | 0 | 0 | — |
case-13 | fail→pass | 11,856 | 4,971 | -58% | 1 | 1 | 0% | 1,991 | 1,493 | -25% | 0 | 0 | — |
case-14 | fail→fail | 28,401 | 27,550 | -3% | 1 | 1 | 0% | 6,171 | 6,783 | +10% | 0 | 0 | — |
case-15 | fail→pass | 27,069 | 15,890 | -41% | 1 | 1 | 0% | 6,074 | 4,215 | -31% | 0 | 0 | — |
case-16 | pass→pass | 30,855 | 26,284 | -15% | 1 | 1 | 0% | 6,183 | 6,386 | +3% | 0 | 0 | — |
case-17 | fail→pass | 28,361 | 23,507 | -17% | 1 | 1 | 0% | 6,119 | 5,504 | -10% | 0 | 0 | — |
case-18 | fail→pass | 13,575 | 15,675 | +15% | 1 | 1 | 0% | 2,049 | 3,555 | +73% | 0 | 0 | — |
case-19 | fail→pass | 9,803 | 4,713 | -52% | 1 | 1 | 0% | 1,519 | 1,346 | -11% | 0 | 0 | — |
case-20 | fail→fail | 10,876 | 4,762 | -56% | 1 | 1 | 0% | 1,536 | 1,324 | -14% | 0 | 0 | — |
case-21 | fail→pass | 29,522 | 27,312 | -7% | 1 | 1 | 0% | 6,166 | 6,778 | +10% | 0 | 0 | — |
case-22 | fail→pass | 30,300 | 30,607 | +1% | 1 | 1 | 0% | 6,169 | 6,781 | +10% | 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 +55 percentage points is the difference between those two pass rates over the 22 comparable cases.
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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/3/2026 | +30% |
Other measured skills in the registry, with their headline benchmark lift.