agent-framework-azure-ai-py
by thegovind
DevOps & CloudBrief
Build Azure AI Foundry agents.
Description
What This Skill Does
Build persistent agents on Azure AI Foundry using the Microsoft Agent Framework Python SDK. Supports function tools, hosted tools like code interpreter and Bing web search, MCP server integration, conversation threads, streaming responses, and structured Pydantic outputs.
Azure AI Foundry manages agent persistence and tool execution server-side, so you don't need to maintain your own state or tool infrastructure.
When to Use It
- Running multi-turn conversations with persistent thread state
- Executing Python code inside an agent to analyze or transform data
- Searching the web via Bing within an automated agent workflow
- Returning typed structured outputs from agent queries using Pydantic models
- Connecting agents to MCP servers for extended tool capabilities
Example Workflow
User asks: Search for Python best practices and give me a structured summary
- 1
Initialize AzureAIAgentsProvider with Azure credentials
- 2
Create agent with HostedWebSearchTool and an AnalysisResult Pydantic response format
- 3
Run agent with user query on a new conversation thread
- 4
Await response and validate against AnalysisResult model
- 5
Return structured object with summary, key_findings, and confidence
AnalysisResult object with summary string, key_findings list, and confidence float
Requirements
- Azure subscription with AI Foundry project access
- AZURE_AI_PROJECT_ENDPOINT environment variable
- AZURE_AI_MODEL_DEPLOYMENT_NAME environment variable
- BING_CONNECTION_ID environment variable (required only for web search)
Installation
clawhub install thegovind/agent-framework-azure-ai-pyRun the CLI command, or paste the prompt into Claude Code, Cursor, Codex or whichever coding agent you use.