Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Frontend UI design system selector and implementation guide covering Glassmorphism, Liquid Glass (Apple-style), Neubrutalism, and Claymorphism. Use when building UI components, choosing a visual aesthetic, implementing design tokens, or auditing accessibility/contrast on themed surfaces. Provides per-style tokens, component patterns, dark mode, and a11y constraints.
.claude/skills/telagod-applying-ui-design-system/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 10% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 10% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -4% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -15% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -22% | 0% |
> 判断先于执行：决定「是否做 / 选什么 / 如何取舍」（栈、方案、架构、权衡）前，先读领域判断内核 skills/_kernel/frontend/SKILL.md——它管 judgment，本秘典管 execution；冲突时以内核判断为准。
> 不同设计语言不可混搭。先选语言，再按 reference 落地组件。
| 想要的感觉 | 选 | 加载 | |---|---|---| | 半透明、磨砂、轻盈、纵深感弱 | Glassmorphism | references/glassmorphism.md | | Apple 风格、立体玻璃、环境光响应、弹簧动效 | Liquid Glass | references/liquid-glass.md | | 粗框硬影、高饱和、纯平、反精致 | Neubrutalism | references/neubrutalism.md | | 软糯黏土、大圆角、双内影、温暖 | Claymorphism | references/claymorphism.md |
每个 reference 自包含「核心要素 + CSS Tokens + 关键约束 + 组件 + 暗色模式 + 无障碍」全套。
无论选哪种风格，下列约束全部适用：
:focus-visible — 必须可见；轮廓与底色有足够反差。键盘可达性是底线。prefers-reduced-motion: reduce — 禁用弹簧、滑入、parallax 等装饰动画；保留功能性 transition。prefers-contrast: more — 减少透明、增强边界、平化阴影。--bg-surface），不要硬编码 #fff/#000 后再交换。tokens-<style>.css 到项目，按需扩展backdrop-filter 元素 ≤ 3（Glassmorphism / Liquid Glass 限制）@supports 或 @media (hover: none) 降级为纯色| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 30,038 | 21,039 | -30% | 1 | 1 | 0% | 4,405 | 4,849 | +10% | 0 | 0 | — |
case-02 | fail→pass | 25,038 | 20,454 | -18% | 1 | 1 | 0% | 4,184 | 4,599 | +10% | 0 | 0 | — |
case-16 | pass→pass | 19,370 | 16,819 | -13% | 1 | 1 | 0% | 2,941 | 3,379 | +15% | 0 | 0 | — |
case-03 | fail→pass | 29,833 | 21,006 | -30% | 1 | 1 | 0% | 4,859 | 4,655 | -4% | 0 | 0 | — |
case-04 | pass→pass | 15,641 | 7,193 | -54% | 1 | 1 | 0% | 2,179 | 1,848 | -15% | 0 | 0 | — |
case-05 | pass→pass | 17,812 | 15,351 | -14% | 1 | 1 | 0% | 2,685 | 3,228 | +20% | 0 | 0 | — |
case-06 | pass→pass | 19,641 | 10,523 | -46% | 1 | 1 | 0% | 2,808 | 2,321 | -17% | 0 | 0 | — |
case-07 | fail→pass | 17,591 | 7,899 | -55% | 1 | 1 | 0% | 2,483 | 2,110 | -15% | 0 | 0 | — |
case-08 | fail→pass | 17,307 | 7,146 | -59% | 1 | 1 | 0% | 2,595 | 2,023 | -22% | 0 | 0 | — |
case-09 | pass→pass | 20,835 | 19,385 | -7% | 1 | 1 | 0% | 3,096 | 3,789 | +22% | 0 | 0 | — |
case-10 | pass→pass | 16,618 | 20,075 | +21% | 1 | 1 | 0% | 2,499 | 3,902 | +56% | 0 | 0 | — |
case-11 | pass→pass | 9,644 | 9,831 | +2% | 1 | 1 | 0% | 1,690 | 2,469 | +46% | 0 | 0 | — |
case-12 | pass→pass | 17,708 | 17,410 | -2% | 1 | 1 | 0% | 2,698 | 3,441 | +28% | 0 | 0 | — |
case-13 | pass→pass | 18,945 | 10,788 | -43% | 1 | 1 | 0% | 2,362 | 2,640 | +12% | 0 | 0 | — |
case-14 | pass→pass | 14,414 | 8,360 | -42% | 1 | 1 | 0% | 1,859 | 2,228 | +20% | 0 | 0 | — |
case-15 | pass→pass | 14,648 | 16,281 | +11% | 1 | 1 | 0% | 2,411 | 3,250 | +35% | 0 | 0 | — |
case-17 | pass→pass | 18,343 | 17,365 | -5% | 1 | 1 | 0% | 2,439 | 3,525 | +45% | 0 | 0 | — |
case-18 | pass→pass | 17,883 | 17,712 | -1% | 1 | 1 | 0% | 2,471 | 3,298 | +33% | 0 | 0 | — |
case-19 | pass→pass | 15,569 | 11,117 | -29% | 1 | 1 | 0% | 2,196 | 2,514 | +14% | 0 | 0 | — |
case-20 | pass→pass | 15,475 | 2,967 | -81% | 1 | 1 | 0% | 2,132 | 1,317 | -38% | 0 | 0 | — |
case-21 | pass→pass | 15,164 | 9,316 | -39% | 1 | 1 | 0% | 2,083 | 2,434 | +17% | 0 | 0 | — |
case-22 | pass→pass | 15,873 | 9,875 | -38% | 1 | 1 | 0% | 2,290 | 2,270 | -1% | 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 +23 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.
Other measured skills in the registry, with their headline benchmark lift.