Official read-only MCP server for HackerOne bug bounty reports, programs, and remediation data.
Storybook
Summary
Official addon that connects a running Storybook to AI agents for component-aware UI generation and testing.
Features
- Exposes a project's own component library and usage guidelines to AI agents
- Returns live story URLs from a running Storybook for visual preview
- Generates and previews new stories for agent-written UI
- Runs targeted interaction and accessibility tests on changed stories only
- Works with any MCP client: Claude Code, Cursor, Gemini CLI, VS Code Copilot, and more
Installation
Set up this MCP server in your favorite AI agent — copy a ready-made configuration below.
Description
The Storybook MCP addon turns a project's own component library into context an AI agent can use directly, instead of asking the agent to guess at conventions from scratch. Once installed and running, it exposes the agent's client to two tools: one that hands over UI development instructions and documented usage guidelines for the project's existing components, and one that returns live story URLs from the running Storybook instance so generated UI can be previewed against real, rendered components.
Because the agent can see what components already exist and how they're meant to be used, it reuses them instead of inventing new one-off markup, then writes matching stories so a human can visually verify each new UI state. From there, Storybook can run interaction and accessibility tests against just the stories that changed, giving a fast, targeted feedback loop rather than a full test-suite run on every edit.
It currently supports Vite-based Storybook setups (React, Next.js, and SvelteKit are called out explicitly) and works with any MCP-compatible client, including Claude Code, Cursor, Gemini CLI, and VS Code Copilot. The addon and its AI features are in preview and currently focused on React projects.
Related MCP Servers
A proxy that lets local-only MCP clients connect to remote servers, handling OAuth and transport differences.
Control Blender from an AI client — build scenes, apply materials, and run Python against the live session.
Run AI-generated code safely in an isolated cloud sandbox, with results returned as data rather than text.