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 Support dark mode with prefers-color-scheme. Check the rendered layout across breakpoints and interaction states before proposing a fix.
.claude/skills/thedaviddias-dark-mode-css/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | 26% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | -16% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | -18% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | -14% | 0% |
| case-05 | ✓→✓ | ＝ Same ✓ | 21% | 0% |
More than half of users prefer dark mode for reduced eye strain, especially in low-light environments. Users with photosensitivity rely on it. Implementing dark mode via prefers-color-scheme respects the user's OS preference without them needing to find a toggle, and CSS custom properties make it a clean, maintainable addition rather than a disruptive refactor.
Check if this CSS supports dark mode. Look for prefers-color-scheme usage or a data-theme attribute pattern. Identify any hard-coded colors that would look wrong in dark mode.
Implement dark mode by extracting colors to CSS custom properties and redefining them inside a prefers-color-scheme: dark media query.
Explain how to implement dark mode with CSS custom properties, the prefers-color-scheme media query, and a JavaScript toggle for user preference.
Review stylesheets, component styles, and responsive states related to Support dark mode with prefers-color-scheme. 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/dark-mode-css
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 14,902 | 11,252 | -24% | 1 | 1 | 0% | 3,440 | 2,894 | -16% | 0 | 0 | — |
case-02 | pass→pass | 7,677 | 5,042 | -34% | 1 | 1 | 0% | 1,736 | 1,431 | -18% | 0 | 0 | — |
case-03 | fail→fail | 4,613 | 4,713 | +2% | 1 | 1 | 0% | 993 | 1,350 | +36% | 0 | 0 | — |
case-04 | pass→pass | 12,004 | 9,105 | -24% | 1 | 1 | 0% | 2,889 | 2,489 | -14% | 0 | 0 | — |
case-05 | pass→pass | 10,105 | 8,131 | -20% | 1 | 1 | 0% | 1,977 | 2,399 | +21% | 0 | 0 | — |
case-06 | pass→pass | 15,250 | 12,973 | -15% | 1 | 1 | 0% | 3,320 | 3,146 | -5% | 0 | 0 | — |
case-07 | pass→pass | 8,944 | 5,521 | -38% | 1 | 1 | 0% | 2,029 | 1,508 | -26% | 0 | 0 | — |
case-08 | fail→pass | 8,041 | 6,837 | -15% | 1 | 1 | 0% | 1,553 | 1,958 | +26% | 0 | 0 | — |
case-09 | pass→pass | 9,511 | 8,598 | -10% | 1 | 1 | 0% | 2,311 | 2,284 | -1% | 0 | 0 | — |
case-10 | pass→pass | 7,634 | 7,265 | -5% | 1 | 1 | 0% | 1,757 | 1,949 | +11% | 0 | 0 | — |
case-11 | pass→pass | 6,942 | 4,473 | -36% | 1 | 1 | 0% | 1,423 | 1,348 | -5% | 0 | 0 | — |
case-12 | pass→pass | 4,426 | 7,028 | +59% | 1 | 1 | 0% | 1,084 | 1,894 | +75% | 0 | 0 | — |
case-13 | pass→pass | 10,831 | 8,064 | -26% | 1 | 1 | 0% | 2,320 | 2,323 | +0% | 0 | 0 | — |
case-14 | pass→pass | 10,620 | 9,043 | -15% | 1 | 1 | 0% | 2,242 | 2,513 | +12% | 0 | 0 | — |
case-15 | pass→pass | 8,987 | 10,956 | +22% | 1 | 1 | 0% | 1,919 | 2,872 | +50% | 0 | 0 | — |
case-16 | pass→pass | 12,966 | 6,688 | -48% | 1 | 1 | 0% | 2,923 | 1,733 | -41% | 0 | 0 | — |
case-17 | pass→pass | 7,787 | 4,698 | -40% | 1 | 1 | 0% | 1,644 | 1,372 | -17% | 0 | 0 | — |
case-18 | pass→pass | 5,452 | 5,233 | -4% | 1 | 1 | 0% | 1,162 | 1,470 | +27% | 0 | 0 | — |
case-19 | pass→pass | 14,022 | 12,769 | -9% | 1 | 1 | 0% | 2,753 | 2,572 | -7% | 0 | 0 | — |
case-20 | pass→pass | 10,807 | 8,494 | -21% | 1 | 1 | 0% | 2,291 | 2,192 | -4% | 0 | 0 | — |
case-21 | pass→pass | 10,154 | 8,312 | -18% | 1 | 1 | 0% | 1,919 | 2,031 | +6% | 0 | 0 | — |
case-22 | pass→pass | 10,295 | 6,845 | -34% | 1 | 1 | 0% | 2,049 | 1,657 | -19% | 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 +5 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.