Build a Microsoft Agent

• Key points are in the visual diagram above.

AutoGen shines in conversational teams. It fails when you forget termination and cost.

A time-boxed workshop: 8 turns, a scribe, a critic, then you leave the room.

Scenario

Task: 'plot y=x^2'. Coder writes matplotlib in the proxy. Critic checks the file exists. Stop.

Solution

In Microsoft Agent Framework, apply Build a Microsoft Agent to this scenario: Task: 'plot y=x^2'. Identify the inputs, run the technique, validate the output, and note one thing you would monitor in production.

Do this before moving to the next module — reading alone is not enough.

Open the Code Walkthrough below and run it locally. Change one parameter related to Build a Microsoft Agent (e.g. model, temperature, top_k, or tool name), observe the difference in output, and write 2–3 sentences explaining what changed.

Highlighted lines show where Build a Microsoft Agent happens in the code.

1user = UserProxyAgent(2    "user",3    human_input_mode="NEVER",4    code_execution_config={"work_dir": "scratch", "use_docker": True},5)6# initiate_chat with max_round=8 and a critic that emits TERMINATE

• use_docker False on untrusted code
• Infinite critic loop

Quick recap — the most important points from this module.