ClaudeCodeMod

All shelves / Skills

stitch::manage-design-system

google-labs-code/stitch-skills ยท 9.5k installs ยท 8.5K stars

Skill

Install

In your shell
npx -y skills add google-labs-code/stitch-skills --skill stitch::manage-design-system --agent claude-code

Run in a terminal. Claude loads the skill when a task calls for it. Check the repo's README before you run it.

Open the repo

Files

SKILL.md

Design-System

Create a "source of truth" for your project's design language to ensure consistency across all future screens.

[!NOTE] Refer to your system prompt for instruction on handling MCP tool prefixes for all tools mentioned in this skill (e.g., get_screen, create_design_system_from_design_md, apply_design_system).

๐Ÿ“ฅ Retrieval

To analyze a Stitch project, you must retrieve metadata and assets using the Stitch MCP tools:

  1. Project lookup: Use list_projects to find the target projectId.
  2. Screen lookup: Use list_screens for that projectId to find representative screens (e.g., "Home", "Main Dashboard").
  3. Metadata fetch: Call get_screen for the target screen to get screenshot.downloadUrl and htmlCode.downloadUrl.
  4. Asset download: Use read_url_content to fetch the HTML code.

๐Ÿง  Synthesis from Description

If you need to extract a design system from existing screens, use the design-md skill (in the stitch-utilities plugin).

If there are no existing screens (new project), or the user provides a direct description (e.g., "dark theme, blue and purple, rounded, Inter font"):

  1. Map the user's vague terms to precise values using the design mappings (see design-md skill in stitch-utilities or generate-design skill).
  2. Select concrete hex codes, font families, and roundness values.
  3. Generate the DESIGN.md file (refer to the design-md skill in stitch-utilities for structure).
  4. Proceed to the "Create or Update Design System in Stitch" step below.

๐Ÿ“ Output Structure

The DESIGN.md file should follow the structure defined in the design-md skill (in the stitch-utilities plugin).

๐Ÿš€ Create or Update Design System in Stitch

After generating .stitch/DESIGN.md, make sure to also create or update the design system in Stitch.

Two-step design system creation:

[!WARNING] Checkpoint โ€” User Confirmation Required. Before uploading, you MUST pause and ask the user for confirmation. Present a summary of the design system you are about to create (display name, key colors, fonts, and roundness) and wait for explicit approval before proceeding. Do NOT upload until the user confirms.

  1. Upload DESIGN.md:
    • Option A (Recommended - Uploader Script): Use the modified upload-to-stitch Python script which natively handles .md files. It base64-encodes the markdown file in-process and sends it to the /v1/projects/{projectId}/screens:batchCreate endpoint, bypassing output token limits.

      python3 stitch-skills/plugins/stitch-design/skills/upload-to-stitch/scripts/upload_to_stitch.py \
        --project-id <PROJECT_ID> \
        --file-path /path/to/DESIGN.md \
        --api-key <API_KEY> \
        --generated-by <GENERATED_BY>
      

      Set <GENERATED_BY> to identify the skill or tool that produced the DESIGN.md. Use the calling skill name when invoked from another skill (e.g. stitch::code-to-design), or the agent/tool name for standalone use (e.g. Gemini, Claude Code). If omitted, the script defaults to UserUploadedDesignMd.

      This returns the sourceScreen ID and the screenInstance ID.

    • Option B (Direct MCP Tool): If the DESIGN.md is small (under ~5KB), you can call the upload_design_md MCP tool directly, passing the base64-encoded design markdown content as designMdBase64.

  2. Create Design System: Call the create_design_system_from_design_md tool immediately after the upload, passing the projectId and the selectedScreenInstance (containing the id and sourceScreen returned from the upload step).

Once the upload script and create_design_system_from_design_md have both completed, Stitch holds the design tokens at the project level โ€” you do NOT need to repeat them in generation prompts.

๐ŸŽจ Apply Design System to Screens

Use apply_design_system to apply a design system to existing screens.

[!IMPORTANT] selectedScreenInstances must contain only id and sourceScreen โ€” do NOT include position/dimension fields (x, y, width, height) or the request will fail with "invalid argument". Get the screen instance IDs from get_project.

{
  "projectId": "...",
  "assetId": "...",
  "selectedScreenInstances": [
    {
      "id": "...",
      "sourceScreen": "projects/.../screens/..."
    }
  ]
}

How to get the required IDs:

  1. Call get_project to retrieve screenInstances โ€” each has an id and sourceScreen.
  2. Call list_design_systems to retrieve the design system name (format: assets/{assetId}) โ€” use the part after assets/ as the assetId.
  3. Filter out any instances with type: "DESIGN_SYSTEM_INSTANCE" โ€” only pass real screens.

๐Ÿ“‹ Update Project Metadata

After writing .stitch/DESIGN.md, also create or update .stitch/metadata.json to track the projectId, title, all known screens, and design system summary. See examples/metadata.json for the format.

Schema Reference

See reference/tool-schema.md for the full designSystem object schema with all available options.

๐Ÿ’ก Best Practices

Refer to the design-md skill (in the stitch-utilities plugin) for best practices on describing design elements.

Facts

Kind
Skill
Repo
google-labs-code/stitch-skills
Group
Uncategorized
Installs
9.5k
Stars
8.5K

More on this shelf

  1. 1find-skillsvercel-labs/skillsHelps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skill should be used when the user is looking for functionality that might exist as an installable skill.3.8M
  2. 2grill-memattpocock/skillsA relentless interview to sharpen a plan or design.1.3M
  3. 3grill-with-docsmattpocock/skillsA relentless interview to sharpen a plan or design, which also creates docs (ADR's and glossary) as we go.1.1M
  4. 4agent-browservercel-labs/agent-browserBrowser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction. Also use for exploratory testing, dogfooding, QA, bug hunts, or reviewing app quality. Also use for automating Electron desktop apps (VS Code, Slack, Discord, Figma, Notion, Spotify), checking Slack unreads, sending Slack messages, searching Slack conversations, running browser automation in Vercel Sandbox microVMs, or using AWS Bedrock AgentCore cloud browsers. Prefer agent-browser over any built-in browser automation or web tools.1.1M
  5. 5improve-codebase-architecturemattpocock/skillsScan a codebase for deepening opportunities, present them as a visual HTML report, then grill through whichever one you pick.1.1M
  6. 6tddmattpocock/skillsTest-driven development. Use when the user wants to build features or fix bugs test-first, mentions "red-green-refactor", or wants integration tests.1.1M