What problem does it solve? Maintaining a consistent visual identity across projects and AI coding agents requires a single machine-readable source of truth for design tokens, but hand-written style guides drift, break WCAG contrast rules, and cannot be consumed programmatically. ## Core Features & Use Cases - Authoring DESIGN.md files: Generate spec-compliant files combining YAML design tokens (colors, typography, spacing, rounded corners, components) with human-readable rationale sections. - Linting and validation: Run the @google/design.md CLI to catch broken token references, WCAG AA contrast failures, out-of-order sections, and orphaned tokens. - Export and diffing: Convert DESIGN.md to Tailwind v3/v4 themes or W3C DTCG JSON, and diff two versions to detect regressions. - Use Case: A team wants every AI-generated UI to follow their brand. They author one DESIGN.md with their palette and typography, lint it for contrast compliance, and export a Tailwind theme so all projects share the same tokens. ## Quick Start Create a DESIGN.md design token spec for my brand with a blue accent color and Inter typography, then lint it and export a Tailwind theme.