Contributions are welcome. This project is meant to be a community resource for anyone who cares about design quality in AI-assisted workflows.
Domain knowledge packaged for reuse. If you're an expert in something design-related and you find yourself repeating the same guidance, it should be a skill.
Requirements:
- Create a directory in
skills/your-skill-name/ - Include a
SKILL.mdwith YAML frontmatter (name, description, triggers) - Content should be actionable, not theoretical
- Include examples, patterns, and anti-patterns
Specialized reviewers with clear scope and output format.
Requirements:
- Create a markdown file in
agents/your-agent.md - Include YAML frontmatter (name, description, tools, model)
- Define what the agent checks, how it gives feedback, and its output format
Always-on constraints that prevent common quality failures.
Requirements:
- Add to the appropriate directory in
rules/ - Include YAML frontmatter (description, globs, alwaysApply)
- Each rule should be one clear, enforceable statement
Pre-built design systems for common aesthetics.
Requirements:
- Create a directory in
design-templates/your-template/ - Include a
DESIGN.mdwith color tokens, typography, spacing, radius, motion, and personality description
- Fork the repo
- Create a branch:
git checkout -b feat/your-contribution - Make your changes
- Run tests:
npm test - Submit a PR with a clear description of what you added and why
- No AI-generated filler content. Write like a person.
- No buzzwords. Plain language.
- Be specific. "Improve design quality" is too vague. "Detect and flag hardcoded hex values that should be design tokens" is specific.
- Test your contribution. Does it actually help?
Be respectful. Be constructive. Be specific. Design is subjective, but quality isn't. Focus feedback on the work, not the person.