Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when reviewing support UI across multi-page flows such as checkout, onboarding, account recovery, or support journeys. Compare multiple pages in the same variation instead of reviewing a single page alone.
.claude/skills/thedaviddias-consistent-help/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 524% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 53% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -16% | 0% |
| case-21 | ✗→✓ | ▲ Improved | -1% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | -7% | 0% |
Users who need help during a task should not have to hunt for it on every screen. Predictable placement reduces stress and makes support easier to find when a user is already struggling.
Audit repeated help mechanisms across this set of pages. Check whether contact links, chat triggers, help-center links, or support phone numbers appear in the same relative order on each page.
Move repeated help mechanisms into a consistent location and keep their relative order stable across the flow at the same breakpoint.
Explain WCAG Consistent Help, what counts as a repeated help mechanism, and why consistent serial order matters more than pixel-perfect placement.
Review page chrome, support UI, and responsive layouts related to Keep repeated help mechanisms in a consistent location. Flag exact pages or breakpoints where repeated help moves to a different relative position.
For full implementation details, code examples, and framework-specific guidance, see references/rule.md.
Rule page: https://frontendchecklist.io/en/rules/accessibility/consistent-help
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 2,721 | 14,928 | +449% | 1 | 1 | 0% | 503 | 3,140 | +524% | 0 | 0 | — |
case-02 | fail→pass | 11,646 | 17,000 | +46% | 1 | 1 | 0% | 2,715 | 4,151 | +53% | 0 | 0 | — |
case-03 | pass→pass | 11,356 | 8,741 | -23% | 1 | 1 | 0% | 2,185 | 2,026 | -7% | 0 | 0 | — |
case-04 | pass→pass | 12,217 | 9,138 | -25% | 1 | 1 | 0% | 2,104 | 2,250 | +7% | 0 | 0 | — |
case-05 | pass→pass | 10,982 | 8,857 | -19% | 1 | 1 | 0% | 2,173 | 2,119 | -2% | 0 | 0 | — |
case-06 | fail→pass | 7,374 | 5,556 | -25% | 1 | 1 | 0% | 1,496 | 1,250 | -16% | 0 | 0 | — |
case-07 | pass→pass | 10,966 | 6,024 | -45% | 1 | 1 | 0% | 1,878 | 1,434 | -24% | 0 | 0 | — |
case-08 | pass→pass | 14,385 | 11,188 | -22% | 1 | 1 | 0% | 2,355 | 2,616 | +11% | 0 | 0 | — |
case-09 | pass→pass | 12,068 | 9,317 | -23% | 1 | 1 | 0% | 2,019 | 1,918 | -5% | 0 | 0 | — |
case-10 | pass→pass | 11,670 | 10,935 | -6% | 1 | 1 | 0% | 2,281 | 1,955 | -14% | 0 | 0 | — |
case-11 | pass→pass | 8,508 | 6,125 | -28% | 1 | 1 | 0% | 1,489 | 1,579 | +6% | 0 | 0 | — |
case-12 | pass→pass | 11,205 | 12,447 | +11% | 1 | 1 | 0% | 2,028 | 2,449 | +21% | 0 | 0 | — |
case-18 | pass→pass | 4,214 | 3,877 | -8% | 1 | 1 | 0% | 819 | 1,029 | +26% | 0 | 0 | — |
case-13 | fail→fail | 13,007 | 9,934 | -24% | 1 | 1 | 0% | 2,362 | 1,912 | -19% | 0 | 0 | — |
case-14 | pass→pass | 5,785 | 4,870 | -16% | 1 | 1 | 0% | 919 | 1,115 | +21% | 0 | 0 | — |
case-15 | pass→pass | 5,618 | 3,440 | -39% | 1 | 1 | 0% | 1,190 | 1,000 | -16% | 0 | 0 | — |
case-16 | pass→pass | 8,883 | 10,514 | +18% | 1 | 1 | 0% | 1,766 | 2,162 | +22% | 0 | 0 | — |
case-17 | pass→pass | 7,901 | 6,984 | -12% | 1 | 1 | 0% | 1,435 | 1,810 | +26% | 0 | 0 | — |
case-19 | pass→pass | 3,745 | 3,722 | -1% | 1 | 1 | 0% | 728 | 1,008 | +38% | 0 | 0 | — |
case-20 | pass→pass | 11,836 | 11,748 | -1% | 1 | 1 | 0% | 1,934 | 2,263 | +17% | 0 | 0 | — |
case-21 | fail→pass | 17,233 | 14,952 | -13% | 1 | 1 | 0% | 3,010 | 2,980 | -1% | 0 | 0 | — |
case-22 | pass→pass | 9,876 | 9,067 | -8% | 1 | 1 | 0% | 1,971 | 2,085 | +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 +18 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.