agentleFS
Joyboy-dyJoyboy-dy/felicio-ai-skills/skills/ux-ui-audit/SKILL.md

ux-ui-audit

Skill0 starsChanged 3 months agoMITView on GitHub

Use when auditing an existing interface for usability, accessibility, responsive behavior, hierarchy, consistency, interaction states, form/table/dialog issues, design-token misuse, or evidence-based UX/UI defects and prioritizing fixes.

3 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

8 rules. 0 are common, 5 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 5

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use frontend-aesthetics after the audit when the user wants a new visual direction.

    line 14
  2. Distinctive:

    Do not report standards compliance that was not actually tested.

    line 23
  3. Very specific:

    Read references/state-patterns.md for state coverage.

    line 58
  4. Distinctive:

    Check semantics, accessible names, heading/landmark structure, keyboard order/operability, focus visibility/management, labels/errors, …

    line 62
  5. Distinctive:

    Use automated tools already present in the repository when available, but manually verify high-impact flows.

    line 64
  6. Distinctive:

    Ask whether the interface makes the next action obvious, groups related content, uses consistent terminology, provides feedback, …

    line 68
  7. Very specific:

    Use assets/audit-report-template.md for a formal report and references/audit-checklist.md before finalizing.

    line 85
  8. Distinctive:

    Prefer a prioritized findings table plus a short summary of systemic patterns.

    line 100

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

UX/UI Audit

Audit what exists, record evidence, prioritize impact, and avoid silently redesigning the product.

Use when

Activate for “audit/review this UI,” accessibility review, responsive/state consistency checks, design-system misuse, interaction friction, contrast/focus problems, or requests for prioritized UI findings.

Use frontend-aesthetics after the audit when the user wants a new visual direction. Use chromatic-design for detailed color-system repair and form-ux-validation for deep form behavior.

Audit stance

  • Separate observed evidence from inference and recommendation.
  • Evaluate the user's task and state transitions, not only screenshots at rest.
  • Severity reflects user impact, frequency/reach, and task blockage—not personal taste.
  • Respect product context and existing design system.
  • Automated/static scans are triage only; runtime and assistive-technology checks are stronger evidence where available.
  • Do not report standards compliance that was not actually tested.

Workflow

1. Establish scope

Identify target pages/components, supported themes/viewports, interaction paths, accessibility target if known, and whether source code, rendered UI, screenshots, or all three are available.

2. Discover design and component conventions

Inspect tokens, primitives, layouts, representative screens, and accessibility utilities. This distinguishes a local defect from an intentional system rule.

3. Run static triage when source is available

python scripts/scan_a11y_static.py <project-root> --json

Investigate matches manually. The script intentionally avoids pretending to be a full accessibility engine.

4. Audit by user journey and state

For each relevant task, inspect:

  • default/content state;
  • loading/skeleton;
  • empty/no-results;
  • error/retry;
  • hover/focus/active/selected/disabled;
  • validation/success where forms exist;
  • overlays/dialogs/menus;
  • long/overflowing content;
  • narrow and wide responsive states;
  • theme variants when supported.

Read references/state-patterns.md for state coverage.

5. Accessibility review

Check semantics, accessible names, heading/landmark structure, keyboard order/operability, focus visibility/management, labels/errors, status announcements, color contrast, non-color cues, target sizes where applicable, reduced motion, and zoom/reflow.

Use automated tools already present in the repository when available, but manually verify high-impact flows. Do not claim a WCAG conformance level from a static source scan.

6. Usability and hierarchy

Ask whether the interface makes the next action obvious, groups related content, uses consistent terminology, provides feedback, prevents/recover from errors, and avoids unnecessary cognitive load. Flag visual inconsistencies when they impair hierarchy or learnability, not merely because two values differ.

7. Responsive behavior

Audit reflow strategy, overflow, tables/data density, nav/action reachability, touch targets, fixed/sticky elements, dialogs, text wrapping, and input ergonomics. Test actual breakpoints/containers rather than only desktop/mobile screenshots.

8. Prioritize findings

Use severity:

  • Critical: blocks a core task, creates serious access/safety/data-loss risk, or makes key content unusable.
  • High: major friction or accessibility failure affecting an important path.
  • Medium: meaningful inconsistency/friction with workaround.
  • Low: polish or minor inconsistency with limited user impact.

For each finding include evidence/location, impact, recommendation, and validation method.

Use assets/audit-report-template.md for a formal report and references/audit-checklist.md before finalizing.

Common failure modes

  • Turning an audit into an unsolicited redesign.
  • Listing generic best practices with no evidence/location.
  • Treating lint/static matches as confirmed accessibility failures.
  • Auditing only the happy/default state.
  • Calling every inconsistency “high severity.”
  • Assuming a component library guarantees accessible usage.
  • Giving exact contrast claims without calculating the rendered pair.
  • Ignoring responsive overflow/keyboard/focus because screenshots look good.

Output

Prefer a prioritized findings table plus a short summary of systemic patterns. Clearly mark what was statically inspected, interactively tested, automatically tested, or not testable in the available environment.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.