| name | audit |
| description | Review interface quality and return prioritized findings with severity, impact, and recommended fixes. |
| args | [{"name":"area","description":"The feature or area to audit (optional)","required":false}] |
| user-invokable | true |
Use this skill for review-only requests. Do not fix issues unless the user explicitly asks.
Load
references/design-system.md
references/interaction-states.md
references/structure-architecture.md
references/content-accessibility.md
references/implementation-quality.md
Audit Dimensions
- Visual quality: hierarchy, spacing, typography, color, depth, brand/product fit, generic AI tells.
- Interaction: hover/focus/active/disabled/loading/error/success states, command hierarchy, empty states.
- Accessibility: semantics, labels, contrast, keyboard paths, focus visibility, touch targets.
- Responsive behavior: mobile/tablet/desktop layout, overflow, text wrapping, hidden critical actions.
- Architecture: component boundaries, token usage, duplicated patterns, state ownership.
- Performance: images, animation properties, unnecessary imports, render hotspots, layout shift.
Report Format
Start with findings ordered by severity.
For each finding include:
- Location: file/component/route and line when available.
- Severity: Critical, High, Medium, or Low.
- Impact: why it matters to users or maintainers.
- Recommendation: specific fix direction.
Then include:
- Top 3 fixes to do first.
- Positive patterns worth preserving.
- Test or verification gaps.
Be strict, but avoid noise. Do not list low-value nitpicks when larger usability or accessibility issues exist.