Agent Operator Lab

This repo contains hands-on notebooks and sample agents for operating and connecting agents in Azure AI Foundry from a Control Plane perspective. The focus is practical: create (or point to) a Foundry project, run lightweight inventory/monitoring checks, and connect both MS Agent Framework agents and Hosted Agents.

What’s Included

• 0_setup/1_setup.ipynb: Bootstrap the minimum Foundry resources (Resource Group, AIServices account, Project), discover the Project endpoint/API key, and write a local config file for reuse.
• 1_controlplane/1_foundry_agent_monitoring.ipynb: Control Plane checks with the Foundry SDK (list deployed agents, inspect assets like connections, and understand where quota is managed).
• 1_controlplane/2_ms_agent_framework_connect_foundry.ipynb: Connect MS Agent Framework agents using AzureAIClient, including conversation/thread linkage and tracing verification in the Foundry portal.
• 1_controlplane/3_hosted_agent_connect_foundry.ipynb: Package agents as containers, push to ACR, and register Hosted Agents using the Hosting Adapter (with sample agents under 1_controlplane/1.1_hosted-agent_sdk/).

Prerequisites

• Python 3.12+
• Azure CLI (az) and an Azure account with access to Azure AI Foundry
• For Hosted Agents: Docker and an Azure Container Registry (ACR)

Setup

uv sync --prerelease=allow
source .venv/bin/activate

Copy environment variables:

cp sample.env .env

Update .env values as needed. Common variables used across the notebooks:

• AZURE_TENANT_ID, AZURE_SUBSCRIPTION_ID
• AZURE_AI_MODEL_DEPLOYMENT_NAME (model deployment name in Foundry)
• AZURE_OPENAI_ENDPOINT (Azure OpenAI endpoint)
• AZURE_OPENAI_CHAT_DEPLOYMENT_NAME (used by some notebooks/samples)
• AZURE_CONTAINER_REGISTRY (required for Hosted Agent container builds)
• BING_GROUNDING_CONNECTION_NAME (only if you run the web-search hosted agent scenario)

Note: authentication is typically done via DefaultAzureCredential (e.g., az login). Some flows may also use AZURE_OPENAI_API_KEY depending on the notebook/sample.

Suggested Run Order

1. Run 0_setup/1_setup.ipynb to create or configure your Foundry Project.
2. Run 1_controlplane/1_foundry_agent_monitoring.ipynb for basic fleet/assets checks.
3. Run 1_controlplane/2_ms_agent_framework_connect_foundry.ipynb to connect MS Agent Framework agents and verify tracing.
4. Run 1_controlplane/3_hosted_agent_connect_foundry.ipynb to build/push container images and register Hosted Agents.

Hosted Agent Samples

Sample agents live under 1_controlplane/1.1_hosted-agent_sdk/:

• calculator-agent: LangGraph + Hosting Adapter (simple arithmetic tools)
• msft-docs-agent: MAF-based example
• workflow-agent: concurrent workflow example
• web-search-agent: grounding with Bing Search connection

Security Notes

• Do not commit .env or 0_setup/.foundry_config.json (they can contain secrets like API keys).
• If a key was ever committed, rotate the key in Azure and rewrite Git history before sharing the repo.