What problem does it solve?
Projects often suffer from inconsistent styling, scattered tokens, and divergent component implementations that make UI maintenance slow and error-prone. This Skill centralizes visual decisions, extracts canonical design tokens, and surfaces precise file:line fixes to restore consistency and accelerate development.
Core Features & Use Cases
- Token extraction: Scan CSS, Tailwind, and styled-component patterns to extract colors, typography, spacing, border-radius, shadows, and breakpoints into a structured design-tokens.json and CSS custom properties.
- Visual audit & scoring: Evaluate UI quality across ten dimensions (color, typography, spacing, components, responsiveness, dark mode, animation, accessibility, density, polish) with concrete examples and remediation steps that reference exact files and line numbers.
- Documentation & preview: Generate a DESIGN.md with rationale for decisions and a self-contained HTML preview illustrating components and theme variations for stakeholder review.
- AI slop detection: Identify generic or AI-generated styling patterns (gratuitous gradients, meaningless glass morphism, overused defaults) and recommend targeted changes.
- Use cases: Start a new design system, audit an existing codebase for visual drift, or review PRs that touch themes, tokens, or component styles.
Quick Start
Use the design-system skill to scan the repository, extract design tokens, produce DESIGN.md, and generate an interactive design-preview.html.