Overview

This directory contains a sample implementation of a Foundry Agent SDK Agent-to-Agent (A2A) client and agent server.

Files

• agent.py: Contains the implementation of the Foundry Agent SDK agent and its executor.
• aqgent_executor.py: Contains the implementation of the executor that processes incoming messages and generates responses.
• client.py: Contains the implementation of the A2A client that communicates with the Foundry Agent SDK agent.

Installation

1. Ensure you have Python 3.13 or higher installed.
2. Install UVicorn on Windows for running the agent server:

   powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"

3. If on macOS or Linux, you can install UVicorn using curl:

   curl -LsSf https://astral.sh/uv/install.sh | sh

4. Install the required dependencies using pip:

   uv sync

For information on installing uv, see the UV Installation Guide

Usage

1. Start the Foundry Agent SDK agent server.
2. Run the A2A client to send a message to the agent and receive the response.

Starting the Agent Server

To start the agent server, run the following command in the terminal:

uv run .

Running the A2A Client

To run the A2A client, execute the following command in a separate terminal:

uv run client.py

Resources

A2A Samples for Azure AI Foundry Agent SDK