a2a-hello-agent

# A2A Hello World A minimal implementation of the Agent-to-Agent (A2A) protocol, demonstrating basic agent communication. ## What is A2A? The Agent-to-Agent (A2A) protocol is an open standard that enables different AI agents to communicate with each other in a structured, predictable way. This project provides a simple "hello world" example of two agents exchanging messages using A2A. ## How It Works In this demo: - Agent 1 sends the message "hello agent 2" - Agent 2 responds with "hello back agent 1" ## Key A2A Concepts Demonstrated ### Agent Cards Each agent has an "agent card" that serves as its identity document. The card describes: - Who the agent is (name, description) - What skills it has - What input/output formats it supports - Authentication requirements (none in this demo) Agent cards are published at the `/.well-known/agent.json` endpoint. ### Tasks Communication in A2A happens within the context of "tasks." A task: - Has a unique ID - Contains a session ID (for related tasks) - Tracks the state of the work (submitted, working, completed, etc.) - Contains messages and artifacts ### Messages and Artifacts - **Messages**: Content exchanged between agents (like "hello agent 2") - **Artifacts**: Results produced by agents (like the response "hello back agent 1") ### JSON-RPC 2.0 All A2A communication uses the JSON-RPC 2.0 format for requests and responses, providing a standardized way to make requests and handle responses. ## How to Run This Demo 1. Clone this repository 2. Install dependencies: `pip install -r requirements.txt` 3. Run the application: `python main.py` 4. Open your browser to the URL shown in the console (typically http://localhost:8080) 5. Click the "Start Conversation" button to see the agents communicate ## Code Structure This implementation is simplified to run within a single Flask application: - Agent 1 and Agent 2 are implemented as different routes within the same app - Both agents expose their agent cards at `/.well-known/agent.json` - Agent 2 handles RPC requests at `/agent2/rpc` - Communication between agents is simulated using Flask's test client ## Real-World A2A In a production environment: - Agents would typically run as separate services - They would communicate over HTTP/HTTPS - Authentication would be implemented - They could handle more complex data types (images, structured data) - Multiple agents could work together on complex tasks ## Further Reading For more information on the A2A protocol, see the [official documentation](https://a2a.org). ## License [MIT License](LICENSE)