← Experts

Orchestrate agents that talk to each other

Design multi-agent architectures where agents discover each other via A2A Agent Cards, communicate through standard protocols, and build on MCP for tool integration — the full stack for agent-to-agent collaboration.

1

Expert Install

Copy this to your agent — it will install, configure, and verify everything.

Say to your agent
Read https://clawhub.md/expert/agent-orchestrator.md and set me up as Agent Orchestrator

Works on OpenClaw, Claude Code, Telegram, Feishu, and any agent interface. Your agent reads agent-orchestrator.md and follows the setup steps inside.

2

Talk and handle it

After setup, say these to your agent:

Build agents that discover, communicate, and… Official

  • "Create an Agent Card for my research agent"
  • "Discover agents that can handle PDF analysis"
  • "Send a long-running task to another agent with streaming"

Build MCP servers that give your… Official

  • "Build an MCP server for my REST API"
  • "Add a database read tool to my MCP server"
  • "Create an MCP server from scratch"

Delegate implementation tasks to Claude Code… Official

  • "Delegate the MCP server boilerplate to a coding agent"
  • "Have Codex scaffold the A2A transport layer"
  • "Run this integration test in a background agent"

View setup file agent-orchestrator.md

Agent Orchestrator

Agentic setup file — share this URL with your agent and it will set everything up for you: https://clawhub.md/expert/agent-orchestrator.md

Goal: Orchestrate agents that talk to each other

What you'll have: Design multi-agent architectures where agents discover each other via A2A Agent Cards, communicate through standard protocols, and build on MCP for tool integration — the full stack for agent-to-agent collaboration.


Step 1: Install

clawhub install openclaw/a2a anthropics/mcp-builder openclaw/coding-agent

Step 2: Configure

Each skill may need credentials or auth before it can act on your behalf.

openclaw/a2a

Build agents that discover, communicate, and collaborate using the A2A v1.0 protocol — Agent Cards, skill querying, and task streaming.

  • Install the A2A SDK: pip install a2a-sdk (Python) or npm 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 implementation tasks to Claude Code or Codex while you focus on orchestration logic and agent architecture.

  • 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/codex then codex 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:

openclaw/a2a

  • "Create an Agent Card for my research agent"
  • "Discover agents that can handle PDF analysis"
  • "Send a long-running task to another agent with streaming"

anthropics/mcp-builder

  • "Build an MCP server for my REST API"
  • "Add a database read tool to my MCP server"
  • "Create an MCP server from scratch"

openclaw/coding-agent

  • "Delegate the MCP server boilerplate to a coding agent"
  • "Have Codex scaffold the A2A transport layer"
  • "Run this integration test in a background agent"

Agent Orchestrator · clawhub.md/expert/agent-orchestrator