Client and server examples for the @wardenprotocol/agent-kit SDK.
pnpm installpnpm echoStarts a minimal A2A server on http://localhost:3000 that echoes back messages. Great for testing with the A2A Inspector or any A2A client.
For Docker-based tools (like A2A Inspector), use:
HOST=host.docker.internal pnpm echocp .env.example .env
# Edit .env and set OPENAI_API_KEY
pnpm agentStarts a dual-protocol agent powered by OpenAI, serving both A2A and LangGraph on the same port.
# Start a server first, then:
pnpm test-clientpnpm dev| Script | Description |
|---|---|
pnpm echo |
Start echo agent (no API key needed) |
pnpm agent |
Start OpenAI dual-protocol agent |
pnpm test-client |
Run A2A client test against a running agent |
pnpm dev |
Run SDK usage overview |
examples/
├── server/
│ ├── echo.ts # Minimal echo agent (A2A only)
│ └── agent.ts # OpenAI agent (A2A + LangGraph)
├── client/
│ ├── client-test.ts # A2A client test
│ └── index.ts # SDK usage overview
├── .env.example
└── package.json