Skill Wiki
A11y Audit V1
Accumulated execution feedback for @community/method-a11y-audit version 1.x. Summarizes observed success rates, common failure modes, and confidence adjustments derived from real audit runs.
$ prime install @community/feedback-a11y-audit-v1Projection
Always in _index.xml · the agent never has to ask for this.
A11yAuditV1Feedback [feedback] v1.0.0
Accumulated execution feedback for @community/method-a11y-audit version 1.x. Summarizes observed success rates, common failure modes, and confidence adjustments derived from real audit runs.
Loaded when retrieval picks the atom as adjacent / supporting.
A11yAuditV1Feedback [feedback] v1.0.0
Accumulated execution feedback for @community/method-a11y-audit version 1.x. Summarizes observed success rates, common failure modes, and confidence adjustments derived from real audit runs.
Target
@community/method-a11y-audit
Observations
- Sample Size: 47
- Period: 2025-Q1 to 2026-Q1
- Success Rate: 0.82
Common Failures:
- Step: @community/tool-axe-core
- Failure: axe-core not available in agent sandbox — falls back to visual-only analysis
- Frequency: 0.31
- Impact: reduces coverage from ~57% to ~15% automated; increases false-negative rate
- Mitigation: always check @community/scope-frontend-review preconditions before starting
- Step: merge-deduplicate violations
- Failure: duplicate violations from axe-core and Lighthouse reported twice, inflating violation count
- Frequency: 0.18
- Impact: user confusion — reported violations > actual violations
- Mitigation: deduplicate by (element-selector, criterion-id) before reporting
- Step: focus-order check
- Failure: DOM order does not reflect visual order due to CSS Flexbox/Grid reordering — false negatives on focus order
- Frequency: 0.24
- Impact: SC 2.4.3 failures missed in automated pass
- Mitigation: flag any use of
order:CSS property orflex-direction: row-reversefor manual review
- Confidence Adjustments:
- @Community/Check Contrast Aa: 0.95
- @Community/Check Focus Visible: 0.87
- @Community/Check Keyboard Reachable: 0.71
Ai Confidence
0.79
Loaded when retrieval picks the atom as a focal / direct hit.
A11yAuditV1Feedback [feedback] v1.0.0
Accumulated execution feedback for @community/method-a11y-audit version 1.x. Summarizes observed success rates, common failure modes, and confidence adjustments derived from real audit runs.
Target
@community/method-a11y-audit
Observations
- Sample Size: 47
- Period: 2025-Q1 to 2026-Q1
- Success Rate: 0.82
Common Failures:
- Step: @community/tool-axe-core
- Failure: axe-core not available in agent sandbox — falls back to visual-only analysis
- Frequency: 0.31
- Impact: reduces coverage from ~57% to ~15% automated; increases false-negative rate
- Mitigation: always check @community/scope-frontend-review preconditions before starting
- Step: merge-deduplicate violations
- Failure: duplicate violations from axe-core and Lighthouse reported twice, inflating violation count
- Frequency: 0.18
- Impact: user confusion — reported violations > actual violations
- Mitigation: deduplicate by (element-selector, criterion-id) before reporting
- Step: focus-order check
- Failure: DOM order does not reflect visual order due to CSS Flexbox/Grid reordering — false negatives on focus order
- Frequency: 0.24
- Impact: SC 2.4.3 failures missed in automated pass
- Mitigation: flag any use of
order:CSS property orflex-direction: row-reversefor manual review
- Confidence Adjustments:
- @Community/Check Contrast Aa: 0.95
- @Community/Check Focus Visible: 0.87
- @Community/Check Keyboard Reachable: 0.71
Ai Confidence
0.79
Source
prime-system/examples/frontend-design/primes/compiled/@community/feedback-a11y-audit-v1/atom.yaml