A2A Integration for Microsoft Agent Framework

Warnings

• breaking API instability and breaking changes: As a beta package, `agent-framework-a2a`'s API surface is subject to frequent changes without strict adherence to semantic versioning. Existing code might break with minor version updates.
  Fix: Regularly check the `microsoft/agent-framework` GitHub repository for updates and breaking change announcements. Pin exact versions for production use and test thoroughly after updates.
• gotcha Requires `agent-framework` context: This library is designed as an integration for the broader Microsoft Agent Framework. Standalone usage without the full framework context or appropriate environment setup (e.g., Azure AI Studio, Prompt Flow) may be complex or non-functional.
  Fix: Refer to the `microsoft/agent-framework` documentation and examples for proper integration within the framework ecosystem.
• gotcha Limited public documentation: Due to its beta status and integration-focused nature, detailed, standalone documentation and examples specifically for `agent-framework-a2a` may be sparse. Expect to infer usage from the main `agent-framework` repository.
  Fix: Consult the `microsoft/agent-framework` GitHub repository and associated Azure AI Studio documentation for guidance.
• gotcha Import path variability: Given the beta status and potential refactoring, the exact import paths for client and model classes may change in future releases. Always verify against the latest framework examples.
  Fix: Verify import paths with the latest `agent-framework` source code or documentation before deployment.

Install

• pip install agent-framework-a2a Install latest beta

Imports

• AgentToAgentClient

  from agent_framework_a2a.client import AgentToAgentClient

• AgentToAgentMessage

  from agent_framework_a2a.models import AgentToAgentMessage

Quickstart

This quickstart verifies the installation by attempting to import core components and demonstrates creating a conceptual A2A message. Full client instantiation and interaction require specific environment configuration and integration with the broader Microsoft Agent Framework or Azure AI Studio.

import os

print("Attempting to import core A2A components from agent-framework-a2a...")
try:
    from agent_framework_a2a.client import AgentToAgentClient
    from agent_framework_a2a.models import AgentToAgentMessage
    print("Successfully imported AgentToAgentClient and AgentToAgentMessage.")

    # This part is conceptual as actual client instantiation and usage
    # heavily depends on the broader Microsoft Agent Framework context and
    # specific environment setup (e.g., Azure AI Studio, Prompt Flow).
    # The below code demonstrates creating a sample message object.

    sender_id = os.environ.get("A2A_SENDER_ID", "agent_alpha")
    recipient_id = os.environ.get("A2A_RECIPIENT_ID", "agent_beta")
    message_content = "Hello from agent_alpha to agent_beta!"

    # Creating a message object (assuming a simple constructor)
    message = AgentToAgentMessage(
        sender_id=sender_id,
        recipient_id=recipient_id,
        content=message_content,
        message_type="text"
    )
    print(f"Created a sample message: Sender={message.sender_id}, Recipient={message.recipient_id}, Content='{message.content}'")

    # To use AgentToAgentClient, you would typically need to configure it
    # with an endpoint and potentially authentication from your environment.
    # client = AgentToAgentClient(endpoint=os.environ.get("A2A_ENDPOINT", ""), api_key=os.environ.get("A2A_API_KEY", ""))
    # response = client.send_message(message)
    # print(f"Conceptual message sent, response: {response}")

except ImportError as e:
    print(f"Failed to import agent-framework-a2a components. Please ensure the library is correctly installed and its API is stable. Error: {e}")
except Exception as e:
    print(f"An unexpected error occurred during quickstart: {e}")