Agent Framework Dev Day

Agent Framework Dev Day is a full-day, instructor-led series of labs that showcase the Microsoft Agent Framework and adjacent tooling across both .NET and Python ecosystems.

Agenda

• Part 1 – Overview, first access (Lab 0), safety
• Part 2 – Workflow patterns and Model Context Protocol (MCP)
• Part 3 – Retrieval-Augmented Generation (RAG) and Agentic RAG

Prerequisites

Core workshop setup

• Git and Github login
• Visual Studio Code with the recommended extensions listed in VSCode-Extensions.md
• Azure subscription with access to Azure OpenAI and Azure AI Search resources, plus any API keys issued by the instructors
• Stable internet connection for package restores and Azure access

.NET track prerequisites

• .NET 10 SDK (labs rely on .NET 10 features and templates)
• Latest C# Dev Kit or C# extension plus the .NET runtime debugging workload in VS Code
• Follow the environment notes in labs/dotnet/README.md for per-lab debugging tips

Python track prerequisites

• Python 3.13 and PIP (exact version or newer point releases) with pip and venv
• Create and activate a virtual environment under labs/python, then install shared dependencies via pip install -r labs/python/requirements.txt
• Review labs/python/README.md for detailed shell-specific setup steps

Repository layout

• labs/dotnet – C# solutions covering hello-world onboarding through MCP, workflows, and Agentic RAG. Every lab folder contains both /begin (exercise) and /solution (reference) subfolders.
• labs/python – Python equivalents for MCP, workflow, and RAG scenarios, plus notebooks to reinforce concepts.
• tools – Utility projects (for example, LabKey encrypters) that support the hands-on exercises.
• VSCode-Extensions.md – Canonical list of extensions, themes, and settings the instructors will reference.

.NET lab index

Lab	Focus	Supplemental docs
lab0-hello-world	Validate SDK install, run first Agent Framework project, explore solution vs. begin folders	labs/dotnet/lab0-hello-world/README.md
lab1-basic-training	Core agent patterns and incremental exercises (instructions inline in begin/; no README yet)	README not provided
lab2-mcp	Build local/remote MCP servers, run mcp-concepts.ipynb, wire up McpAgentClient	labs/dotnet/lab2-mcp/README.md
lab2-workflow	Implement sequential, concurrent, and human-in-the-loop workflows with Microsoft.Agents.AI.Workflows	labs/dotnet/lab2-workflow/README.md
lab3-agentic-rag	Agentic RAG scenarios with Azure AI Search and Agent Framework orchestration	labs/dotnet/lab3-agentic-rag/begin/README.md

Each .NET lab also includes notebooks (for example, mcp-concepts.ipynb, workflow-concepts.ipynb) that can be launched directly from VS Code for guided exploration.

Python lab index

Lab	Focus	Supplemental docs
lab0-hello-world	Environment smoke test and first Python agent (instructions inside begin/)	README not provided
lab1-safety	Progressive exercises reinforcing Agent Framework primitives (see begin/ folder)	README not provided
lab2-mcp	Python MCP servers/clients plus mcp-concepts.ipynb	labs/python/lab2-mcp/README.md
lab2-workflow	Workflow patterns implemented with Python executors and notebooks	labs/python/lab2-workflow/README.md
lab3-agentic-rag	Agentic RAG build-out leveraging Azure AI Search tooling	labs/python/lab3-agentic-rag/begin/README.md

Additional references

• Keep the VSCode-Extensions.md list close by when setting up your editor profiles.
• Most labs provide EXERCISES.md within the begin/ folder; switch to the solution/ folder whenever you need a working reference.
• Interactive notebooks live alongside the labs (for example, labs/dotnet/lab2-mcp/begin/mcp-concepts.ipynb). Run them in VS Code or Jupyter to validate your environment before coding.

Presented by

• Bill Wilder
• Jason Haley
• Udaiappa Ramachandran

Sponsored by

Date	Location	Sponsors
16-Jan-2026	Burlington	Microsoft, Akumina