Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when reviewing stylesheets, component styles, and responsive behavior related to Use CSS custom properties for design tokens. Check the rendered layout across breakpoints and interaction states before proposing a fix.
.claude/skills/thedaviddias-css-custom-properties/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | -1% | 0% |
| case-21 | ✓→✗ | ▼ Worse | 133% | 0% |
| case-10 | ✓→✓ | ＝ Same ✓ | 45% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | 12% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | 42% | 0% |
Hard-coded color values and magic numbers scattered across CSS files make global changes (a rebrand, a dark mode, a spacing adjustment) require finding and updating dozens of individual lines. Custom properties centralize these values so one change propagates everywhere — and enable runtime theming that preprocessor variables cannot.
Find hard-coded color values, spacing values, and font sizes in this CSS file that should be replaced with CSS custom properties.
Extract repeated and meaningful values into CSS custom properties on :root and replace all usages with var() references.
Explain CSS custom properties: how they differ from preprocessor variables, how cascade and inheritance work, and how to use them for theming.
Review stylesheets, component styles, and responsive states related to Use CSS custom properties for design tokens. Flag exact selectors, declarations, or breakpoints that violate the rule in the rendered UI.
For full implementation details, code examples, and framework-specific guidance, see references/rule.md.
Rule page: https://frontendchecklist.io/en/rules/css/css-custom-properties
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-10 | pass→pass | 4,654 | 2,923 | -37% | 1 | 1 | 0% | 586 | 848 | +45% | 0 | 0 | — |
case-01 | pass→pass | 4,793 | 3,808 | -21% | 1 | 1 | 0% | 1,066 | 1,193 | +12% | 0 | 0 | — |
case-02 | fail→pass | 9,056 | 5,640 | -38% | 1 | 1 | 0% | 1,669 | 1,644 | -1% | 0 | 0 | — |
case-03 | pass→pass | 2,031 | 1,571 | -23% | 1 | 1 | 0% | 405 | 575 | +42% | 0 | 0 | — |
case-04 | pass→pass | 5,382 | 7,469 | +39% | 1 | 1 | 0% | 1,056 | 1,567 | +48% | 0 | 0 | — |
case-05 | pass→pass | 4,375 | 4,762 | +9% | 1 | 1 | 0% | 939 | 1,349 | +44% | 0 | 0 | — |
case-06 | pass→pass | 4,557 | 4,884 | +7% | 1 | 1 | 0% | 1,062 | 1,287 | +21% | 0 | 0 | — |
case-07 | pass→pass | 12,026 | 9,101 | -24% | 1 | 1 | 0% | 2,312 | 2,246 | -3% | 0 | 0 | — |
case-08 | pass→pass | 8,217 | 7,680 | -7% | 1 | 1 | 0% | 1,682 | 1,732 | +3% | 0 | 0 | — |
case-09 | pass→pass | 2,786 | 3,406 | +22% | 1 | 1 | 0% | 610 | 981 | +61% | 0 | 0 | — |
case-11 | pass→pass | 9,178 | 5,578 | -39% | 1 | 1 | 0% | 1,910 | 1,440 | -25% | 0 | 0 | — |
case-12 | pass→pass | 7,767 | 5,083 | -35% | 1 | 1 | 0% | 1,711 | 1,415 | -17% | 0 | 0 | — |
case-13 | pass→pass | 8,982 | 7,906 | -12% | 1 | 1 | 0% | 1,816 | 1,976 | +9% | 0 | 0 | — |
case-14 | pass→pass | 2,969 | 1,862 | -37% | 1 | 1 | 0% | 566 | 583 | +3% | 0 | 0 | — |
case-15 | pass→pass | 1,508 | 1,468 | -3% | 1 | 1 | 0% | 280 | 526 | +88% | 0 | 0 | — |
case-16 | pass→pass | 2,834 | 2,817 | -1% | 1 | 1 | 0% | 672 | 920 | +37% | 0 | 0 | — |
case-17 | pass→pass | 4,826 | 4,263 | -12% | 1 | 1 | 0% | 945 | 1,175 | +24% | 0 | 0 | — |
case-18 | pass→pass | 3,245 | 3,658 | +13% | 1 | 1 | 0% | 733 | 1,033 | +41% | 0 | 0 | — |
case-19 | pass→pass | 10,750 | 8,456 | -21% | 1 | 1 | 0% | 2,329 | 2,211 | -5% | 0 | 0 | — |
case-20 | pass→pass | 2,699 | 5,550 | +106% | 1 | 1 | 0% | 614 | 1,454 | +137% | 0 | 0 | — |
case-21 | pass→fail | 2,590 | 3,471 | +34% | 1 | 1 | 0% | 428 | 997 | +133% | 0 | 0 | — |
case-22 | pass→pass | 3,115 | 4,403 | +41% | 1 | 1 | 0% | 697 | 1,138 | +63% | 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 0 percentage points is the difference between those two pass rates over the 22 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.