google-labs-code/stitch-skills
A library of Agent Skills designed to work with the Stitch MCP server. Each skill follows the Agent Skills open standard, for compatibility with coding agents such as Antigravity, Gemini CLI, Claude Code, Cursor.
What it solves
This project provides a set of specialized skills and plugins for AI coding agents to integrate with Google Stitch. It bridges the gap between design and code by allowing agents to generate designs, convert existing code into designs, and transform those designs into production-ready frontend components (such as React or React Native).
How it works
It implements the Agent Skills open standard, providing a structured set of instructions, scripts, and resources that AI agents (like Codex, Claude Code, Cursor, and OpenCode) can load. These skills interact with the Stitch MCP (Model Context Protocol) server to perform actions such as extracting HTML from web apps, managing design systems via DESIGN.md files, and automating the creation of UI components.
Who it’s for
Developers and designers using AI coding agents who want to automate the workflow between UI design in Google Stitch and frontend implementation.
Highlights
- Design Automation: Generate new screens from text/images or convert frontend code back into Stitch designs.
- Code Generation: Convert Stitch designs into React, React Native, or React + Vite dashboards.
- Design System Management: Extract and apply design tokens and themes using
DESIGN.mdfiles. - Multi-Agent Compatibility: Works with a variety of coding agents including Codex, Claude Code, Cursor, and OpenCode.
- Visual Walkthroughs: Includes a skill to generate walkthrough videos of projects using Remotion.
Related
- Project
- Project
- Project
- Project
- Project