Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when reviewing templates, rendered HTML, or shared components related to Make drag and drop accessible. Validate the final browser-facing markup, not just the source framework abstraction.
.claude/skills/thedaviddias-draggable-accessibility/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | -28% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 11% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | -24% | 0% |
| case-13 | ✓→✓ | ＝ Same ✓ | 5% | 0% |
| case-05 | ✓→✓ | ＝ Same ✓ | -24% | 0% |
Drag and drop is mouse-dependent by default—keyboard and screen reader users are completely locked out without proper alternatives and ARIA announcements.
Verify drag and drop interfaces have keyboard alternatives, proper ARIA attributes, and live region announcements.
Implement keyboard alternatives (arrow keys, Enter/Space), aria-grabbed, aria-dropeffect, and status announcements.
Explain how accessible drag and drop implementations provide equivalent functionality for keyboard and screen reader users.
Review templates, server-rendered HTML, and shared components that output markup related to Make drag and drop accessible. Flag exact elements, attributes, and routes where the rendered HTML violates the rule.
For full implementation details, code examples, and framework-specific guidance, see references/rule.md.
Rule page: https://frontendchecklist.io/en/rules/html/draggable-accessibility
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 19,180 | 14,682 | -23% | 1 | 1 | 0% | 3,768 | 3,331 | -12% | 0 | 0 | — |
case-02 | fail→fail | 11,689 | 8,377 | -28% | 1 | 1 | 0% | 2,138 | 1,777 | -17% | 0 | 0 | — |
case-03 | fail→fail | 12,491 | 9,416 | -25% | 1 | 1 | 0% | 2,321 | 2,130 | -8% | 0 | 0 | — |
case-04 | pass→pass | 11,420 | 5,644 | -51% | 1 | 1 | 0% | 2,064 | 1,569 | -24% | 0 | 0 | — |
case-13 | pass→pass | 5,507 | 3,638 | -34% | 1 | 1 | 0% | 1,063 | 1,113 | +5% | 0 | 0 | — |
case-05 | pass→pass | 8,994 | 4,216 | -53% | 1 | 1 | 0% | 1,620 | 1,231 | -24% | 0 | 0 | — |
case-06 | fail→pass | 10,958 | 6,382 | -42% | 1 | 1 | 0% | 2,077 | 1,492 | -28% | 0 | 0 | — |
case-07 | pass→pass | 14,333 | 9,517 | -34% | 1 | 1 | 0% | 2,469 | 2,000 | -19% | 0 | 0 | — |
case-08 | pass→pass | 10,915 | 9,959 | -9% | 1 | 1 | 0% | 2,077 | 1,949 | -6% | 0 | 0 | — |
case-09 | pass→pass | 12,890 | 9,574 | -26% | 1 | 1 | 0% | 2,247 | 1,872 | -17% | 0 | 0 | — |
case-10 | fail→pass | 6,228 | 5,378 | -14% | 1 | 1 | 0% | 1,148 | 1,276 | +11% | 0 | 0 | — |
case-11 | pass→pass | 7,533 | 5,618 | -25% | 1 | 1 | 0% | 1,414 | 1,381 | -2% | 0 | 0 | — |
case-12 | pass→pass | 9,454 | 4,133 | -56% | 1 | 1 | 0% | 1,850 | 1,063 | -43% | 0 | 0 | — |
case-14 | pass→pass | 7,011 | 3,285 | -53% | 1 | 1 | 0% | 1,309 | 1,048 | -20% | 0 | 0 | — |
case-15 | pass→pass | 10,942 | 7,163 | -35% | 1 | 1 | 0% | 2,112 | 1,386 | -34% | 0 | 0 | — |
case-16 | pass→pass | 4,123 | 3,439 | -17% | 1 | 1 | 0% | 746 | 790 | +6% | 0 | 0 | — |
case-17 | pass→pass | 12,870 | 5,878 | -54% | 1 | 1 | 0% | 2,055 | 1,260 | -39% | 0 | 0 | — |
case-18 | pass→pass | 12,214 | 6,814 | -44% | 1 | 1 | 0% | 2,167 | 1,447 | -33% | 0 | 0 | — |
case-19 | pass→pass | 6,217 | 5,728 | -8% | 1 | 1 | 0% | 1,160 | 1,314 | +13% | 0 | 0 | — |
case-20 | pass→pass | 16,302 | 12,328 | -24% | 1 | 1 | 0% | 3,060 | 2,939 | -4% | 0 | 0 | — |
case-21 | pass→pass | 7,325 | 4,709 | -36% | 1 | 1 | 0% | 1,332 | 1,324 | -1% | 0 | 0 | — |
case-22 | pass→pass | 7,038 | 5,198 | -26% | 1 | 1 | 0% | 1,515 | 1,334 | -12% | 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 +9 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.