About design-critic-skill
design-critic-skill guides specialists in reviewing screenshots, Figma frames, or source context and turning observable visual issues into actionable fixes. It distinguishes evidence types and includes privacy and validation steps.
Key features
- 01Screenshot, Figma, and code critique modes
- 02Actionable findings tied to visible issues
- 03Critique-to-implementation fix briefs
- 04Privacy and evidence-type checks
- 05Structured critique report template
Use cases
- 01Reviewing a screenshot for visual hierarchy issues
- 02Assessing a Figma frame before implementation
- 03Turning critique findings into a webapp fix brief
Install
Copy the command and run it in a terminal.
$
npx skills add https:/ / github.com/ sergekostenchuk/ ui-ux-agent-skill-system --skill design-critic-skillAfter installing, type/design-critic-skillin Claude Code to confirm it loaded.
Bundled files
References and scripts installed along with SKILL.md.
agents/openai.yamlassets/critique-report.template.mdevals/evals.jsonreferences/anti-slop-patterns.mdhierarchy-rubric.mdreference-products.mdtaste-constraints.mdDetails
From the project READMEDesign Critic Skill
Use this specialist when the main need is visual quality judgment and actionable polish, not full implementation or automated audit.
Operating Modes
screenshot-critique: review a screenshot or image.figma-critique: review a provided Figma frame/context.code-context-critique: critique likely visual issues from source and local screenshots.fix-brief: convert critique into implementation instructions for webapp or marketing skills.
Required Workflow
- Read references/hierarchy-rubric.md.
- Read references/anti-slop-patterns.md before calling a design "polished".
- Use references/taste-constraints.md to keep critique observable.
- Apply
../../shared/privacy-policy.mdbefore Figma/screenshots/external tools. - Use assets/critique-report.template.md for structured output.
- Mark evidence type: screenshot, Figma, source, browser, or
Manual.
Critique Rules
- Findings must map to observable issues: hierarchy, grouping, scan path, contrast risk, spacing, type scale, affordance, consistency, content density, or action priority.
- Avoid vague taste claims without fix direction.
- Do not invent accessibility, DOM, performance, or analytics evidence.
- Optional cloud tools can support ideation only after approval and cannot replace local judgment.
Validation
python3 $CODEX_SKILLS_DIR/.system/skill-creator/scripts/quick_validate.py $CODEX_SKILLS_DIR/design-critic-skill
python3 $CODEX_SKILLS_DIR/senior-skill-architect/scripts/lint_production_skill.py $CODEX_SKILLS_DIR/design-critic-skill
Output Shape
Input:
Evidence type:
Top findings:
Fix brief:
What not to change:
Verification:
Privacy notes:
Comments
Markdown supported