agent_with_a2a_config

A2A client examples (agent_with_a2a_config, agent_with_a2a_client)

Outbound A2A: your agent calls remote A2A agents as tools.

• agent_with_a2a_config — agent.WithA2AConfig(agent.A2AServers{<serverName>: cfg}); SDK builds the default pkg/a2a/client per server.
• agent_with_a2a_client — a2aclient.NewClient(...) + WithA2AClients; same env and testing flow — see ../agent_with_a2a_client/README.md.

Prerequisites

• Temporal + LLM in examples/.env (see ../env.sample).

Required for A2A: A2A_URL — base URL of the remote agent (scheme + host + port, no path). Optional: A2A_SERVER_NAME, A2A_TOKEN, A2A_HEADERS (JSON), A2A_TIMEOUT_SECONDS, A2A_SKIP_TLS_VERIFY (dev), A2A_ALLOW_SKILLS / A2A_BLOCK_SKILLS (comma-separated; mutually exclusive).

Run

From examples/:

go run ./agent_with_a2a_config
go run ./agent_with_a2a_config "What tools do you have available?"

go run ./agent_with_a2a_client
go run ./agent_with_a2a_client "What tools do you have available?"

Testing against a real A2A server

There is no fixed demo URL in this repo—run an A2A-compatible HTTP server locally (or use your deployment), set A2A_URL to its base URL. The SDK loads the agent card from the well-known path (same as a2aproject/a2a-go a2asrv.WellKnownAgentCardPath) and uses the JSON-RPC endpoint advertised on the card.

Worked example — Python helloworld (a2a-samples)

git clone https://github.com/a2aproject/a2a-samples
cd a2a-samples/samples/python/agents/helloworld
uv run .   # server on port 9999 by default

In another terminal:

curl -sS "http://localhost:9999/.well-known/agent-card.json" | head

In examples/.env set A2A_URL=http://localhost:9999 (or http://127.0.0.1:9999), then from examples/:

go run ./agent_with_a2a_config

Other sample servers

Quick sanity check

Replace host/port if needed:

curl -sS "http://localhost:9999/.well-known/agent-card.json" | head

Env vars (A2A client)

See A2A_* rows in examples/README.md and ../env.sample.