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 containment to limit repaint scope. Check the rendered layout across breakpoints and interaction states before proposing a fix.
.claude/skills/thedaviddias-css-containment/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-17 | ✗→✓ | ▲ Improved | -38% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | -5% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | -11% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | 50% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 16% | 0% |
By default, a CSS change anywhere on the page can potentially affect any other element's layout. CSS containment declares that certain elements are isolated — changes inside them don't affect the outside world. This lets browsers skip large portions of the rendering pipeline for unchanged content, which is especially valuable for pages with many independent components.
Identify components in this CSS that are visually self-contained (no overflow, no interaction with elements outside) and could benefit from CSS containment.
Add appropriate contain values to self-contained components. Add content-visibility: auto to repeated off-screen list items.
Explain CSS containment values, how they limit rendering scope, and when content-visibility: auto provides the biggest benefit.
Review stylesheets, component styles, and responsive states related to Use CSS containment to limit repaint scope. 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-containment
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 6,870 | 4,310 | -37% | 1 | 1 | 0% | 1,325 | 1,265 | -5% | 0 | 0 | — |
case-02 | pass→pass | 7,156 | 5,481 | -23% | 1 | 1 | 0% | 1,545 | 1,371 | -11% | 0 | 0 | — |
case-03 | pass→pass | 2,946 | 2,928 | -1% | 1 | 1 | 0% | 595 | 893 | +50% | 0 | 0 | — |
case-04 | pass→pass | 8,260 | 7,935 | -4% | 1 | 1 | 0% | 1,719 | 1,993 | +16% | 0 | 0 | — |
case-05 | pass→pass | 3,938 | 6,366 | +62% | 1 | 1 | 0% | 763 | 1,649 | +116% | 0 | 0 | — |
case-06 | pass→pass | 7,390 | 4,326 | -41% | 1 | 1 | 0% | 1,570 | 1,228 | -22% | 0 | 0 | — |
case-07 | pass→pass | 5,918 | 4,003 | -32% | 1 | 1 | 0% | 1,168 | 1,166 | -0% | 0 | 0 | — |
case-08 | pass→pass | 10,913 | 7,052 | -35% | 1 | 1 | 0% | 2,064 | 1,681 | -19% | 0 | 0 | — |
case-09 | pass→pass | 3,773 | 4,122 | +9% | 1 | 1 | 0% | 720 | 1,167 | +62% | 0 | 0 | — |
case-10 | pass→pass | 8,591 | 7,599 | -12% | 1 | 1 | 0% | 1,747 | 1,890 | +8% | 0 | 0 | — |
case-11 | pass→pass | 7,268 | 7,502 | +3% | 1 | 1 | 0% | 1,462 | 1,762 | +21% | 0 | 0 | — |
case-12 | pass→pass | 11,866 | 6,997 | -41% | 1 | 1 | 0% | 2,089 | 1,713 | -18% | 0 | 0 | — |
case-13 | pass→pass | 11,494 | 8,096 | -30% | 1 | 1 | 0% | 2,121 | 1,936 | -9% | 0 | 0 | — |
case-14 | pass→pass | 7,292 | 5,519 | -24% | 1 | 1 | 0% | 1,469 | 1,466 | -0% | 0 | 0 | — |
case-15 | pass→pass | 12,818 | 10,254 | -20% | 1 | 1 | 0% | 2,240 | 2,251 | +0% | 0 | 0 | — |
case-16 | pass→pass | 11,904 | 8,064 | -32% | 1 | 1 | 0% | 2,172 | 2,060 | -5% | 0 | 0 | — |
case-17 | fail→pass | 9,127 | 4,182 | -54% | 1 | 1 | 0% | 2,083 | 1,282 | -38% | 0 | 0 | — |
case-18 | pass→pass | 6,155 | 5,255 | -15% | 1 | 1 | 0% | 1,248 | 1,304 | +4% | 0 | 0 | — |
case-19 | pass→pass | 4,167 | 3,729 | -11% | 1 | 1 | 0% | 884 | 1,031 | +17% | 0 | 0 | — |
case-20 | pass→pass | 3,496 | 6,376 | +82% | 1 | 1 | 0% | 618 | 1,530 | +148% | 0 | 0 | — |
case-21 | pass→pass | 6,767 | 7,265 | +7% | 1 | 1 | 0% | 1,604 | 1,921 | +20% | 0 | 0 | — |
case-22 | pass→pass | 10,322 | 8,455 | -18% | 1 | 1 | 0% | 1,981 | 2,100 | +6% | 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.