design-lab

Generates five distinct UI design variants with interactive feedback collection and implementation plans.

3|Updated Apr 20, 2026
One-click install
npx skills add https://github.com/langgenius/due-date-hq-jwl --skill design-lab-langgenius
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: design-lab
Source: https://github.com/langgenius/due-date-hq-jwl/tree/main/.claude/skills/design-lab
Command: npx skills add https://github.com/langgenius/due-date-hq-jwl --skill design-lab-langgenius

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Choosing a UI direction without seeing concrete options leads to costly rework and subjective debates. This Skill runs a structured design exploration workflow that produces five meaningfully different UI variants you can compare side-by-side in your own app before committing to one. ## Core Features & Use Cases - Guided Design Interview: Collects scope, pain points, brand tone, persona, and constraints, then detects your framework (Next.js, Vite, Remix, Nuxt, Astro), package manager, and styling system (Tailwind, MUI, Chakra, Ant Design) automatically. - Five-Variant Design Lab: Generates a temporary route (e.g., /__design_lab) rendering five variants that each explore a different axis—hierarchy, layout, density, interaction model, and brand expression—using your project's existing design tokens. - Interactive Feedback Overlay: A Figma-like overlay lets you click any element in the browser, attach comments, and paste structured feedback back into the conversation for targeted refinement. - Synthesis and Finalization: Combines liked elements into a hybrid variant, shows a final preview, then cleans up all temporary files and produces DESIGN_PLAN.md plus a DESIGN_MEMORY.md for future sessions. - Use Case: Redesigning a checkout summary component—run the interview, review five variants in your running dev server, pick a winner with tweaks, and receive a step-by-step implementation plan with accessibility and testing checklists. ## Quick Start Ask the AI to run the design-lab skill to explore redesign options for a specific component or page in your project.

Frequently Asked Questions about design-lab

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate multiple UI design variants for a component?

Run the design-lab skill with a target component or page. It interviews you about scope, brand, and constraints, then creates a temporary route rendering five variants side-by-side—each exploring hierarchy, layout, density, interaction, or brand expression—using your project's existing design tokens.

What frameworks does the design lab support?

It detects Next.js (App and Pages Router), Vite, Remix, Nuxt, and Astro from config files, and identifies styling systems including Tailwind, Material UI, Chakra UI, Ant Design, styled-components, Emotion, and CSS Modules. If detection fails, it asks you directly.

How does the interactive feedback overlay work?

The lab page includes a FeedbackOverlay component rendered alongside the variants. You click Add Feedback, select any element, type a comment, and submit; the overlay produces structured Markdown with CSS selectors that you paste back into the terminal for the AI to apply.

Does the design lab leave temporary files in my project?

No. After you confirm a final design—or abort at any point—it deletes the .claude-design/ directory and all temporary route files it created. It only writes DESIGN_PLAN.md and DESIGN_MEMORY.md as permanent outputs.

Can it reuse my existing design system instead of generic styles?

Yes. It reads your Tailwind config, CSS variables, or UI library theme, and scans existing components to infer colors, spacing, radius, and typography. All five variants use your project's visual language rather than predefined styles.

What happens if I cancel the design exploration midway?

Saying cancel, abort, or stop triggers immediate cleanup of all temporary lab and route files. No implementation plan is generated and Design Memory is not updated, leaving your project exactly as it was.