agentleFS
memi-designmemi-design/design-skills/skills/figma-audit/SKILL.md

figma-audit

Skill7 starsChanged 4 months agoMITView on GitHub

Audit Figma file for design system consistency, accessibility, token adoption, Code Connect coverage

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 3 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Before a design handoff to development

    line 15
  2. Very specific:

    After pulling a design system (memi pull)

    line 16
  3. Distinctive:

    When components look inconsistent or broken

    line 18
  4. Distinctive:

    Before establishing Code Connect mappings

    line 19

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “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

/figma-audit — Design System Audit & Quality Check

Audit a Figma file for design system consistency, accessibility compliance, token adoption, and Code Connect coverage. Produces actionable findings. Requires /figma-use.

Freedom Level: Read-Only + Report

Audit skills read and analyze but do not modify the canvas. Output is a structured report with severity levels.

When to Use

  • Before a design handoff to development
  • After pulling a design system (memi pull)
  • Periodic design system health checks
  • When components look inconsistent or broken
  • Before establishing Code Connect mappings

Audit Checklist

1. Token Adoption

figma_get_variables → list all defined tokens
get_design_context for each page → check usage

Report:
  ✓ Bound to variable: colors/primary/500
  ✗ Raw hex: #3b82f6 (should be colors/primary/500)
  ✗ Magic number: 16px padding (should be spacing/4)

Metric: % of visual properties bound to variables
Target: 90%+

2. Component Consistency

figma_search_components → list all components
figma_get_component_details for each → check structure

Report:
  ✓ Auto Layout applied
  ✓ Properties documented
  ✓ Variants complete (default, hover, active, disabled, focused)
  ✗ Missing state: Button has no "loading" variant
  ✗ Inconsistent padding: Card uses 16px, CardHeader uses 12px

Metric: % of components with all required states
Target: 100% for atoms, 90% for molecules

3. Atomic Design Compliance

Check each component's classification:
  ✓ Atom: no nested component instances (correct)
  ✗ Atom: Button contains Icon instance (should be molecule or use instance swap)
  ✓ Molecule: FormField composes Label + Input + HelpText
  ✗ Organism: LoginForm classified as molecule (too complex)

Metric: % of components correctly classified
Target: 95%+

4. Accessibility

For each component:
  ✓ Contrast ratio: text on background meets WCAG AA (4.5:1)
  ✗ Touch target: Button is 32px tall (needs 44px minimum)
  ✓ Focus indicator: visible focus ring
  ✗ Missing: no aria-label on icon-only button

For each page:
  ✓ Heading hierarchy: h1 → h2 → h3 (no skips)
  ✗ Missing landmark: no main content area defined
  ✓ Color not sole indicator: error states use icon + color

Metric: WCAG violations per page
Target: 0 critical, <5 minor

5. Code Connect Coverage

get_code_connect_map → current mappings

Report:
  ✓ Mapped: Button → src/components/ui/button.tsx
  ✓ Mapped: Card → src/components/ui/card.tsx
  ✗ Unmapped: MetricCard (no code counterpart)
  ✗ Stale: FormField maps to deleted file

Metric: % of components with valid Code Connect
Target: 80%+ for atoms, 60%+ for molecules

6. Naming Conventions

Check against naming rules:
  ✓ PascalCase: MetricCard, LoginForm
  ✗ Wrong case: metric-card (should be MetricCard)
  ✗ Vague name: "Frame 47" (rename to meaningful name)
  ✓ Token path: colors/primary/500

Metric: % of elements following naming conventions
Target: 100%

7. Auto Layout Health

For each container:
  ✓ Has Auto Layout
  ✗ Uses absolute positioning (should be Auto Layout)
  ✓ Sizing: fill container (correct for content area)
  ✗ Sizing: hug contents (should be fill for responsive element)
  ✓ Spacing uses token values

Metric: % of containers using Auto Layout
Target: 95%+

Output Format

╔══════════════════════════════════════╗
║     Design System Audit Report       ║
╚══════════════════════════════════════╝

Score: 82/100

Token Adoption:      87% ✓
Component Quality:   79% ⚠
Atomic Compliance:   95% ✓
Accessibility:       74% ⚠
Code Connect:        65% ✗
Naming:             92% ✓
Auto Layout:         89% ✓

Critical Issues (fix immediately):
  1. Button has no focus indicator
  2. 23 raw hex values should be variables
  3. FormField Code Connect points to deleted file

Warnings (fix soon):
  1. MetricCard missing "loading" variant
  2. 5 components have inconsistent padding

Recommendations:
  1. Run: memi tokens → re-export and bind
  2. Run: add_code_connect_map for 12 unmapped components
  3. Add missing variant states to 8 components

Integration

  • Save audit results to .memoire/audit.json for tracking over time
  • Compare against previous audits to show improvement
  • Block memi sync if critical issues exceed threshold

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.