Build and ship AI agents end to end
Scaffold Claude API apps with the Agent SDK (Python/TS), build MCP servers that give LLMs new tools, add A2A protocol for agent-to-agent communication, publish skills to the Claude Code marketplace, and delegate coding to agents — the full agent builder stack.
Expert Install
Copy this to your agent — it will install, configure, and verify everything.
Works on OpenClaw, Claude Code, Telegram, Feishu, and any agent interface. Your agent reads ai-builder-expert.md and follows the setup steps inside.
Talk and handle it
After setup, say these to your agent:
Build with the Claude Agent SDK… Official
- › "Build an agent with the Claude Agent SDK"
- › "Add tool use and structured output to my agent"
- › "Set up subagents for parallel task execution"
Build MCP servers that let LLMs… Official
- › "Build an MCP server for my REST API"
- › "Add Elicitation support for user prompts"
- › "Create a new MCP server with TypeScript SDK"
Add A2A v1.0 support — generate… Official
- › "Generate an Agent Card for my agent"
- › "Add A2A skill discovery to my agent"
- › "Set up task streaming between my agent and others"
Package workflows as skills and publish… Official
- › "Package this agent as a publishable skill"
- › "Audit my SKILL.md for marketplace requirements"
- › "Create a namespaced plugin for the Claude Code marketplace"
Delegate boilerplate, test generation, and refactors… Official
- › "Delegate writing tests for this module to an agent"
- › "Have Claude Code scaffold the boilerplate for this feature"
▶ View setup file ai-builder-expert.md
Agent Builder
Agentic setup file — share this URL with your agent and it will set everything up for you:
https://clawhub.md/expert/ai-builder-expert.md
Goal: Build and ship AI agents end to end
What you'll have: Scaffold Claude API apps with the Agent SDK (Python/TS), build MCP servers that give LLMs new tools, add A2A protocol for agent-to-agent communication, publish skills to the Claude Code marketplace, and delegate coding to agents — the full agent builder stack.
Step 1: Install
clawhub install anthropics/claude-api anthropics/mcp-builder openclaw/a2a anthropics/skill-creator openclaw/coding-agent
Step 2: Configure
Each skill may need credentials or auth before it can act on your behalf.
openclaw/a2a
Add A2A v1.0 support — generate Agent Cards, enable inter-agent discovery and task streaming, so your agent can collaborate with other agents.
- Install the A2A SDK:
pip install a2a-sdk(Python) ornpm install @a2a/sdk(TypeScript) - No additional API keys required — A2A is a local protocol
- For remote agent discovery, deploy your Agent Card to a public endpoint
openclaw/coding-agent
Delegate boilerplate, test generation, and refactors to Claude Code or Codex — so you stay in flow on architecture and product.
- Ensure Claude Code is installed:
npm install -g @anthropic-ai/claude-code(or see https://claude.ai/code) - Optional — for Codex delegation:
npm install -g @openai/codexthencodex login - No extra environment variables needed if Claude Code is already working in your session
Step 3: Try it
After setup, say these to your agent to verify everything works:
anthropics/claude-api
- "Build an agent with the Claude Agent SDK"
- "Add tool use and structured output to my agent"
- "Set up subagents for parallel task execution"
anthropics/mcp-builder
- "Build an MCP server for my REST API"
- "Add Elicitation support for user prompts"
- "Create a new MCP server with TypeScript SDK"
openclaw/a2a
- "Generate an Agent Card for my agent"
- "Add A2A skill discovery to my agent"
- "Set up task streaming between my agent and others"
anthropics/skill-creator
- "Package this agent as a publishable skill"
- "Audit my SKILL.md for marketplace requirements"
- "Create a namespaced plugin for the Claude Code marketplace"
openclaw/coding-agent
- "Delegate writing tests for this module to an agent"
- "Have Claude Code scaffold the boilerplate for this feature"
Agent Builder · clawhub.md/expert/ai-builder-expert