What problem does it solve? Building UI without a documented design rationale leads to inconsistent components, hardcoded values, and generic-looking interfaces that drift across sessions. This Skill enforces traceable design decisions by capturing user intent, exploring the product domain, and persisting a reusable design system in .interface-design/system.md and system.json. ## Core Features & Use Cases - Design System Generation: Runs intent questions, domain exploration, and direction selection to produce system.md plus a machine-readable system.json with tokens, motion, and patterns. - Component Construction & Improvement: Builds new components with mandatory per-component checkpoints, improves existing UI in place, or extracts implicit design patterns from existing code into a proposed system. - Craft Validation: Applies four tests (Swap, Squint, Signature, Token) to verify every visual value traces back to a design token before persisting results. - Use Case: A developer starting a new SaaS dashboard runs init to define a calm-authority direction, then builds each component against saved tokens so future sessions stay visually consistent. ## Quick Start Ask the AI to run the design skill in init mode to explore intent and generate a design system for your project.