Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Define or refine a project design system for web and mobile interfaces. Use after visual direction is known and before broad component implementation to establish tokens, type, spacing, layout, states, accessibility, and responsive component rules.
.claude/skills/qierkang-ui-ux-pro-max-design-system/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | -1% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 176% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -42% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 5% | 0% |
| case-15 | ✗→✓ | ▲ Improved | -62% | 0% |
Use this adaptation for design-system decisions, not as a second competing art director.
Update docs/design/DESIGN.md and the owning UI specification with semantic color, typography, spacing, radius, elevation, motion, grid, breakpoints, density, component states, accessibility, reduced motion, and data-display rules.
Do not run upstream installers, create symlinks, or overwrite provider configuration.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 5,912 | 21,665 | +266% | 1 | 1 | 0% | 329 | 340 | +3% | 0 | 0 | — |
case-02 | pass→pass | 42,272 | 36,993 | -12% | 1 | 1 | 0% | 6,701 | 6,805 | +2% | 0 | 0 | — |
case-03 | fail→pass | 40,551 | 41,756 | +3% | 1 | 1 | 0% | 8,238 | 8,175 | -1% | 0 | 0 | — |
case-04 | pass→fail | 15,537 | 20,313 | +31% | 1 | 1 | 0% | 2,765 | 4,229 | +53% | 0 | 0 | — |
case-05 | pass→pass | 43,609 | 34,450 | -21% | 1 | 1 | 0% | 3,344 | 6,551 | +96% | 0 | 0 | — |
case-06 | pass→pass | 29,572 | 39,210 | +33% | 1 | 1 | 0% | 6,113 | 8,376 | +37% | 0 | 0 | — |
case-07 | fail→fail | 10,438 | 24,356 | +133% | 1 | 1 | 0% | 2,001 | 4,110 | +105% | 0 | 0 | — |
case-08 | fail→pass | 4,307 | 9,647 | +124% | 1 | 1 | 0% | 719 | 1,986 | +176% | 0 | 0 | — |
case-09 | fail→pass | 31,332 | 16,982 | -46% | 1 | 1 | 0% | 5,602 | 3,255 | -42% | 0 | 0 | — |
case-10 | pass→pass | 15,046 | 15,535 | +3% | 1 | 1 | 0% | 2,249 | 2,798 | +24% | 0 | 0 | — |
case-11 | fail→pass | 13,367 | 12,419 | -7% | 1 | 1 | 0% | 2,170 | 2,271 | +5% | 0 | 0 | — |
case-12 | pass→pass | 9,417 | 9,346 | -1% | 1 | 1 | 0% | 1,567 | 1,928 | +23% | 0 | 0 | — |
case-13 | fail→fail | 4,451 | 6,244 | +40% | 1 | 1 | 0% | 238 | 552 | +132% | 0 | 0 | — |
case-14 | fail→fail | 16,777 | 28,293 | +69% | 1 | 1 | 0% | 2,862 | 5,895 | +106% | 0 | 0 | — |
case-15 | fail→pass | 11,651 | 4,730 | -59% | 1 | 1 | 0% | 1,880 | 717 | -62% | 0 | 0 | — |
case-16 | fail→pass | 10,908 | 7,430 | -32% | 1 | 1 | 0% | 1,983 | 1,452 | -27% | 0 | 0 | — |
case-17 | pass→pass | 10,512 | 11,247 | +7% | 1 | 1 | 0% | 1,809 | 2,217 | +23% | 0 | 0 | — |
case-18 | fail→fail | 4,432 | 8,664 | +95% | 1 | 1 | 0% | 193 | 570 | +195% | 0 | 0 | — |
case-19 | fail→fail | 9,476 | 5,784 | -39% | 1 | 1 | 0% | 1,599 | 481 | -70% | 0 | 0 | — |
case-20 | fail→fail | 6,244 | 8,357 | +34% | 1 | 1 | 0% | 1,049 | 1,679 | +60% | 0 | 0 | — |
case-21 | fail→fail | 21,882 | 26,326 | +20% | 1 | 1 | 0% | 3,775 | 4,981 | +32% | 0 | 0 | — |
case-22 | fail→fail | 4,664 | 5,589 | +20% | 1 | 1 | 0% | 177 | 486 | +175% | 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 17 counted toward the lift figure. The other 5 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 +23 percentage points is the difference between those two pass rates over the 17 comparable cases. 1 case got worse with the skill loaded, and it is 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.