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 Avoid embedded and inline CSS. Check the rendered layout across breakpoints and interaction states before proposing a fix.
.claude/skills/thedaviddias-embedded-or-inline-css/SKILL.md| Model | Eval pass | Runs |
|---|---|---|
| gemini-3.6-flash | 100% | 26 |
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | -6% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | -7% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | -24% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | -24% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | -40% | 0% |
Inline CSS cannot be cached separately, increases HTML payload, breaks separation of concerns, and is harder to maintain in large codebases.
Verify that inline CSS and embedded styles are avoided except for valid use cases like critical CSS, dynamic styles, or performance optimization.
Move inline and embedded CSS to external stylesheets, keeping only critical above-the-fold CSS inline for performance optimization.
Explain why external CSS is preferred for maintainability, caching, and separation of concerns, with exceptions for critical CSS and dynamic styles.
Review stylesheets, component styles, and responsive states related to Avoid embedded and inline CSS. 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/embedded-or-inline-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 | 12,526 | 10,075 | -20% | 1 | 1 | 0% | 2,278 | 2,121 | -7% | 0 | 0 | — |
case-02 | pass→pass | 13,974 | 9,798 | -30% | 1 | 1 | 0% | 2,581 | 1,967 | -24% | 0 | 0 | — |
case-03 | pass→pass | 11,739 | 7,329 | -38% | 1 | 1 | 0% | 2,141 | 1,626 | -24% | 0 | 0 | — |
case-04 | pass→pass | 8,105 | 4,471 | -45% | 1 | 1 | 0% | 1,949 | 1,179 | -40% | 0 | 0 | — |
case-05 | pass→pass | 17,599 | 12,117 | -31% | 1 | 1 | 0% | 2,602 | 2,418 | -7% | 0 | 0 | — |
case-06 | pass→pass | 12,519 | 8,731 | -30% | 1 | 1 | 0% | 2,547 | 1,941 | -24% | 0 | 0 | — |
case-07 | pass→pass | 14,618 | 8,971 | -39% | 1 | 1 | 0% | 2,635 | 1,938 | -26% | 0 | 0 | — |
case-08 | fail→pass | 15,310 | 13,458 | -12% | 1 | 1 | 0% | 2,968 | 2,786 | -6% | 0 | 0 | — |
case-09 | pass→pass | 9,006 | 4,052 | -55% | 1 | 1 | 0% | 1,706 | 1,095 | -36% | 0 | 0 | — |
case-10 | pass→pass | 8,265 | 7,363 | -11% | 1 | 1 | 0% | 1,691 | 1,602 | -5% | 0 | 0 | — |
case-11 | pass→pass | 12,863 | 8,417 | -35% | 1 | 1 | 0% | 2,090 | 1,653 | -21% | 0 | 0 | — |
case-12 | pass→pass | 18,043 | 14,934 | -17% | 1 | 1 | 0% | 2,753 | 2,619 | -5% | 0 | 0 | — |
case-13 | pass→pass | 5,423 | 5,876 | +8% | 1 | 1 | 0% | 1,061 | 1,357 | +28% | 0 | 0 | — |
case-14 | pass→pass | 12,223 | 5,684 | -53% | 1 | 1 | 0% | 2,300 | 1,367 | -41% | 0 | 0 | — |
case-15 | pass→pass | 10,977 | 9,291 | -15% | 1 | 1 | 0% | 1,896 | 1,756 | -7% | 0 | 0 | — |
case-16 | pass→pass | 10,688 | 7,611 | -29% | 1 | 1 | 0% | 1,764 | 1,735 | -2% | 0 | 0 | — |
case-17 | pass→pass | 14,101 | 10,243 | -27% | 1 | 1 | 0% | 2,550 | 2,109 | -17% | 0 | 0 | — |
case-18 | pass→pass | 13,393 | 5,973 | -55% | 1 | 1 | 0% | 2,771 | 1,373 | -50% | 0 | 0 | — |
case-19 | pass→pass | 15,027 | 13,383 | -11% | 1 | 1 | 0% | 2,752 | 2,467 | -10% | 0 | 0 | — |
case-20 | pass→pass | 8,389 | 7,662 | -9% | 1 | 1 | 0% | 1,538 | 1,716 | +12% | 0 | 0 | — |
case-21 | pass→pass | 7,918 | 6,487 | -18% | 1 | 1 | 0% | 1,337 | 1,359 | +2% | 0 | 0 | — |
case-22 | pass→pass | 13,348 | 12,211 | -9% | 1 | 1 | 0% | 2,496 | 2,443 | -2% | 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.