agentleFS
podopodo/design-agent-skills/skills/extract-design-md/SKILL.md

extract-design-md

Skill32 starsChanged 6 months agoMITView on GitHub

Reverse-engineer a Stitch-compatible DESIGN.md design system from existing frontend source code. Analyzes component files, stylesheets, Tailwind configs, theme definitions, and design tokens in React, Vue, Svelte, Angular, or plain HTML/CSS, and writes a structured document that can be version-controlled, shared with the team, and used to keep future UI consistent.

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

Files
SKILL.md

extract-design-md

Catalogue stub — full skill: google-labs-code/stitch-skills

<!-- das:install:start — generated from stub.yaml by `npm run generate`; edit stub.yaml, not this block -->

Decision tree

  1. Is the full skill already installed? Upstream names this skill stitch-extract-design-md, so it installs next to this pointer (the pointer is not replaced).
    • Global: f=~/.agents/skills/stitch-extract-design-md/SKILL.md; [ -f "$f" ] && ! grep -q "^das:" "$f" && echo "installed" || echo "pointer"
    • Project: f=.agents/skills/stitch-extract-design-md/SKILL.md; [ -f "$f" ] && ! grep -q "^das:" "$f" && echo "installed" || echo "pointer"
    • installed → the skill is installed as stitch-extract-design-md, invoke it and proceed
    • pointer → go to step 2
  2. Detect scope, then install:

    [ -e ~/.agents/skills/extract-design-md ] && echo "global" || echo "project"

    Global:

    npx skills add google-labs-code/stitch-skills --skill stitch::extract-design-md -g -y

    Project:

    npx skills add google-labs-code/stitch-skills --skill stitch::extract-design-md -y

    Claude Code: send a command as a chat message starting with ! to run it without leaving the conversation.

Heads-up

  • Requires: Optional Stitch MCP server, only to push the resulting DESIGN.md into a Stitch project (the extraction itself reads source files only)
  • License: Apache-2.0
<!-- das:install:end -->

Invoke after install

  • Skill name: stitch-extract-design-md (the upstream name is stitch::extract-design-md; the CLI installs it under the sanitised directory name)
  • Trigger phrases: "extract design md", "design system from source code", "reverse engineer design system", "DESIGN.md"

What it does

Reads a frontend codebase and extracts its design system into a DESIGN.md file in the format Google Stitch uses: colors, typography, spacing, component patterns, and layout principles, plus notes for Stitch generation. It works from source files only (stylesheets, component files, Tailwind and theme configs, CSS-in-JS), so it needs no build, no running app, and no Stitch connection. Typical uses: auditing the visual language of a repo, reconciling inconsistent styles, or migrating a project's identity into Stitch. The direction is code to DESIGN.md, not Stitch output to DESIGN.md. Part of Google Labs' official Stitch skills.

When NOT to use

  • You have a rendered Stitch project and want its DESIGN.md — use the design-md skill from google-stitch-skills instead
  • You want design tokens (JSON/CSS) pulled from a live website rather than a DESIGN.md from source — use extract-design-system
  • You only want a DESIGN.md for a well-known brand — use brand-design-md

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.