Can I use local tools with this framework?

Yes, you can pass standard Python functions directly to the agent, which are automatically converted into tools the agent can call during execution.

Does this skill support real-time streaming?

Absolutely, the skill includes implementation patterns for agent.run_stream() to provide immediate feedback to users as the agent generates its response.

What is the main benefit of using the Microsoft Agent Framework on Azure?

It allows for the creation of persistent, stateful agents that can utilize Azure-hosted tools like code interpreters and vector searches while maintaining long-running conversation threads.

How does this handle authentication for Azure services?

It leverages the azure-identity library, supporting both AzureCliCredential for local development and DefaultAzureCredential for secure production environments.

Azure AI Agent Framework (Python)

Name: Azure AI Agent Framework (Python)
Author: microsoft

bymicrosoft

•

1,777

•

クラウドインフラストラクチャ

Builds and manages persistent, tool-enabled AI agents using the Microsoft Agent Framework and Azure AI Foundry.

This skill empowers developers to create robust, production-grade AI agents by leveraging the Microsoft Agent Framework Python SDK and Azure AI Foundry. It provides specialized guidance for implementing persistent conversation threads, integrating hosted tools like code interpreters and Bing search, and connecting Model Context Protocol (MCP) servers. Whether you are building complex multi-tool agents or simple streaming assistants, this skill ensures best practices for authentication, structured outputs, and efficient resource management within the Azure ecosystem.

主な機能

01Persistent conversation thread management with AgentThread

02Streaming response implementation for real-time user interaction

03Type-safe structured outputs using Pydantic models

04Integration of hosted tools including Code Interpreter and Web Search

05Support for Model Context Protocol (MCP) servers (hosted and client-managed)

061,777 GitHub stars

ユースケース

01Building automated data analysis agents that output structured JSON for downstream systems

02Creating research assistants that use Bing search and code execution to verify data

03Developing persistent support bots that maintain context across multiple user sessions

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add microsoft/skills agent-framework-azure-ai-py

For use in Claude.ai and ChatGPT

Download Skill