Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when reviewing sticky navigation, consent banners, chat widgets, skip links, or any fixed-position UI. Validate the actual tab order and viewport behavior at realistic zoom levels.
.claude/skills/thedaviddias-focus-not-obscured/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-16 | ✗→✓ | ▲ Improved | 3% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | -25% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | -11% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | -16% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | 9% | 0% |
A visible focus ring is not enough if the focused element sits behind a sticky header or footer. Keyboard users need to see the actual focused control and the focus indicator at the same time in order to continue the task.
scroll-padding or equivalent spacing when fixed headers or footers are presentAudit keyboard focus with all sticky UI active. Check headers, footers, cookie banners, promo bars, and chat launchers for cases where the focused element scrolls underneath them.
Add layout spacing such as scroll-padding, reduce overlay height, or change focus/scroll behavior so the focused element remains visible when tabbing.
Explain WCAG Focus Not Obscured, why fixed-position UI commonly breaks it, and how scroll-padding and overlay layout choices prevent the issue.
Review sticky layout, focus styles, overlays, and keyboard navigation related to Keep focused elements unobscured. Flag exact selectors, breakpoints, or UI states where focused controls can be covered.
For full implementation details, code examples, and framework-specific guidance, see references/rule.md.
Rule page: https://frontendchecklist.io/en/rules/accessibility/focus-not-obscured
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-04 | pass→pass | 11,724 | 7,896 | -33% | 1 | 1 | 0% | 2,039 | 1,532 | -25% | 0 | 0 | — |
case-02 | pass→pass | 9,589 | 7,293 | -24% | 1 | 1 | 0% | 1,890 | 1,680 | -11% | 0 | 0 | — |
case-03 | pass→pass | 10,739 | 7,399 | -31% | 1 | 1 | 0% | 1,958 | 1,635 | -16% | 0 | 0 | — |
case-01 | pass→pass | 6,034 | 4,279 | -29% | 1 | 1 | 0% | 1,198 | 1,302 | +9% | 0 | 0 | — |
case-05 | pass→pass | 8,292 | 6,035 | -27% | 1 | 1 | 0% | 1,456 | 1,442 | -1% | 0 | 0 | — |
case-06 | pass→pass | 7,014 | 7,402 | +6% | 1 | 1 | 0% | 1,554 | 1,690 | +9% | 0 | 0 | — |
case-07 | pass→pass | 4,022 | 3,893 | -3% | 1 | 1 | 0% | 727 | 1,068 | +47% | 0 | 0 | — |
case-08 | pass→pass | 10,527 | 7,827 | -26% | 1 | 1 | 0% | 1,725 | 1,834 | +6% | 0 | 0 | — |
case-09 | pass→pass | 4,592 | 5,303 | +15% | 1 | 1 | 0% | 888 | 1,364 | +54% | 0 | 0 | — |
case-10 | pass→pass | 7,751 | 5,455 | -30% | 1 | 1 | 0% | 1,458 | 1,491 | +2% | 0 | 0 | — |
case-11 | pass→pass | 4,843 | 3,682 | -24% | 1 | 1 | 0% | 982 | 1,018 | +4% | 0 | 0 | — |
case-12 | pass→pass | 6,696 | 5,721 | -15% | 1 | 1 | 0% | 1,396 | 1,473 | +6% | 0 | 0 | — |
case-13 | pass→pass | 7,027 | 5,709 | -19% | 1 | 1 | 0% | 1,161 | 1,406 | +21% | 0 | 0 | — |
case-14 | pass→pass | 14,387 | 13,148 | -9% | 1 | 1 | 0% | 2,444 | 2,661 | +9% | 0 | 0 | — |
case-15 | pass→pass | 9,147 | 6,834 | -25% | 1 | 1 | 0% | 1,580 | 1,513 | -4% | 0 | 0 | — |
case-16 | fail→pass | 12,906 | 11,810 | -8% | 1 | 1 | 0% | 2,300 | 2,376 | +3% | 0 | 0 | — |
case-17 | pass→pass | 13,030 | 10,051 | -23% | 1 | 1 | 0% | 2,290 | 2,136 | -7% | 0 | 0 | — |
case-18 | pass→pass | 9,392 | 6,050 | -36% | 1 | 1 | 0% | 1,625 | 1,357 | -16% | 0 | 0 | — |
case-19 | pass→pass | 9,919 | 7,772 | -22% | 1 | 1 | 0% | 1,898 | 1,780 | -6% | 0 | 0 | — |
case-20 | pass→pass | 10,118 | 9,038 | -11% | 1 | 1 | 0% | 1,685 | 1,807 | +7% | 0 | 0 | — |
case-21 | pass→pass | 11,445 | 12,772 | +12% | 1 | 1 | 0% | 1,990 | 2,649 | +33% | 0 | 0 | — |
case-22 | pass→pass | 4,172 | 4,982 | +19% | 1 | 1 | 0% | 728 | 1,197 | +64% | 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.