Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Educational agent for WCAG 3.0 (W3C Accessibility Guidelines). Explains methodology changes, outcome-based conformance, the APCA contrast algorithm, functional needs categories, and new cognitive criteria. WCAG 3.0 is in early Working Draft status.
.claude/skills/community-access-wcag-3-0-preview/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -14% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -18% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -19% | 0% |
| case-19 | ✗→✓ | ▲ Improved | -18% | 0% |
| case-16 | ✓→✗ | ▼ Worse | 0% | 0% |
You help teams understand what WCAG 3.0 will require and how to prepare for the transition from WCAG 2.2.
WCAG 3.0 is an early Working Draft and NOT a W3C standard. Do NOT abandon WCAG 2.2 compliance.
For existing audits: read findings, assess WCAG 3.0 severity changes, identify new coverage areas.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 20,678 | 14,635 | -29% | 1 | 1 | 0% | 3,093 | 2,648 | -14% | 0 | 0 | — |
case-02 | pass→pass | 13,465 | 11,265 | -16% | 1 | 1 | 0% | 2,152 | 2,068 | -4% | 0 | 0 | — |
case-03 | pass→pass | 13,801 | 11,261 | -18% | 1 | 1 | 0% | 2,222 | 2,030 | -9% | 0 | 0 | — |
case-04 | pass→pass | 14,587 | 13,587 | -7% | 1 | 1 | 0% | 2,176 | 2,406 | +11% | 0 | 0 | — |
case-05 | pass→pass | 15,888 | 16,440 | +3% | 1 | 1 | 0% | 2,539 | 2,840 | +12% | 0 | 0 | — |
case-06 | pass→pass | 14,346 | 13,529 | -6% | 1 | 1 | 0% | 2,370 | 2,400 | +1% | 0 | 0 | — |
case-07 | pass→pass | 13,184 | 10,981 | -17% | 1 | 1 | 0% | 2,043 | 1,860 | -9% | 0 | 0 | — |
case-08 | pass→pass | 17,010 | 12,705 | -25% | 1 | 1 | 0% | 2,624 | 2,255 | -14% | 0 | 0 | — |
case-09 | fail→pass | 11,590 | 7,327 | -37% | 1 | 1 | 0% | 1,847 | 1,514 | -18% | 0 | 0 | — |
case-10 | fail→fail | 19,810 | 17,796 | -10% | 1 | 1 | 0% | 3,291 | 3,173 | -4% | 0 | 0 | — |
case-11 | pass→pass | 12,033 | 10,650 | -11% | 1 | 1 | 0% | 1,823 | 1,886 | +3% | 0 | 0 | — |
case-12 | fail→pass | 12,781 | 8,725 | -32% | 1 | 1 | 0% | 1,854 | 1,494 | -19% | 0 | 0 | — |
case-13 | pass→pass | 15,567 | 14,016 | -10% | 1 | 1 | 0% | 2,277 | 2,378 | +4% | 0 | 0 | — |
case-14 | pass→pass | 17,134 | 17,211 | +0% | 1 | 1 | 0% | 2,632 | 3,035 | +15% | 0 | 0 | — |
case-15 | pass→pass | 13,700 | 12,234 | -11% | 1 | 1 | 0% | 2,175 | 2,209 | +2% | 0 | 0 | — |
case-16 | pass→fail | 6,196 | 4,776 | -23% | 1 | 1 | 0% | 1,032 | 1,027 | -0% | 0 | 0 | — |
case-17 | pass→pass | 14,360 | 9,592 | -33% | 1 | 1 | 0% | 2,187 | 1,708 | -22% | 0 | 0 | — |
case-18 | pass→pass | 15,992 | 13,047 | -18% | 1 | 1 | 0% | 2,350 | 2,298 | -2% | 0 | 0 | — |
case-19 | fail→pass | 15,588 | 10,766 | -31% | 1 | 1 | 0% | 2,464 | 2,023 | -18% | 0 | 0 | — |
case-20 | pass→pass | 6,100 | 9,002 | +48% | 1 | 1 | 0% | 1,003 | 1,742 | +74% | 0 | 0 | — |
case-21 | pass→pass | 4,970 | 5,610 | +13% | 1 | 1 | 0% | 894 | 1,287 | +44% | 0 | 0 | — |
case-22 | fail→fail | 8,123 | 7,956 | -2% | 1 | 1 | 0% | 1,430 | 1,558 | +9% | 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 +14 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.