techniqueintermediateGeneral AI

Microsoft Agent Framework: Build and Orchestrate Multi-Agent AI Workflows in Python and .NET

Microsoft Agent Framework is an open-source library for building, coordinating, and deploying AI agents that can work together in complex pipelines — available for both Python and .NET developers. It supports graph-based workflows where multiple specialized agents hand off tasks to each other, with built-in features like checkpointing, human-in-the-loop approval steps, and a visual DevUI for debugging. Business owners can use it to automate multi-step processes like customer support escalation, document processing pipelines, or internal research workflows without stitching together separate tools.

AI SETUP PROMPT

Paste into Claude Code or Codex CLI — it will scan your project and set everything up

# Apply Technique: Microsoft Agent Framework: Build and Orchestrate Multi-Agent AI Workflows in Python and .NET

## What This Is
Microsoft Agent Framework is an open-source library for building, coordinating, and deploying AI agents that can work together in complex pipelines — available for both Python and .NET developers. It supports graph-based workflows where multiple specialized agents hand off tasks to each other, with built-in features like checkpointing, human-in-the-loop approval steps, and a visual DevUI for debugging. Business owners can use it to automate multi-step processes like customer support escalation, document processing pipelines, or internal research workflows without stitching together separate tools.

Source: https://github.com/microsoft/agent-framework

## Before You Start

Scan my workspace and analyze:
- The project language, framework, and directory structure
- Existing AI provider config (check .env, .env.local, config files for API keys — OpenRouter, OpenAI, Anthropic, Google AI, etc.)
- Whether this repository or a similar tool is already cloned or installed

Then ask me before proceeding:
1. Which AI provider/API should this use? (Use whatever I already have configured, or ask me to set one up — options include direct provider APIs or a unified service like OpenRouter)
2. Where in my project should this be integrated?
3. Are there any customizations I need (model preferences, naming conventions, constraints)?

## Fetch the Source

Clone or inspect the repository to understand what needs to be installed:
```bash
gh repo clone microsoft/agent-framework
```
Review the README, directory structure, and any install instructions before proceeding.

## What to Implement

This is an **AI Technique** — a pattern or methodology for working with AI models.

- Explain how this technique applies to my current project and what benefit it provides
- Implement it in a way that fits my existing codebase — suggest concrete files to modify or create
- If it requires specific model capabilities (structured output, function calling, etc.), verify my current provider supports them
- Show me a working example I can test immediately

## Additional Context

- Install the framework locally by running `pip install agent-framework --pre` in your terminal and verify it works by importing the package in a Python script.
- Clone the sample workflows directory from the GitHub repo (`python/samples/03-workflows/`) and run one of the pre-built multi-agent examples to see graph-based orchestration in action.
- Launch the DevUI package (`python/packages/devui/`) against your sample agent to visually inspect message flow, debug agent handoffs, and identify where to insert human-in-the-loop checkpoints.

## Reference Implementation

```
pip install agent-framework --pre
# .NET alternative:
# dotnet add package Microsoft.Agents.AI
```

## Guidelines

- Adapt everything to my existing project — do not assume a specific stack or directory layout
- Use whichever AI provider I already have configured; if I need a new one, tell me what to sign up for and I'll give you the key
- Check my .env files for existing API keys (OpenRouter, OpenAI, Anthropic, Google AI) before asking me to add one
- Review any fetched code for safety before installing or executing it
- After setup, run a quick verification and show me a summary of exactly what was installed, where, and how to use it

3,423 charactersCompatible with Claude Code & Codex CLI

MANUAL SETUP STEPS

01Install the framework locally by running `pip install agent-framework --pre` in your terminal and verify it works by importing the package in a Python script.
02Clone the sample workflows directory from the GitHub repo (`python/samples/03-workflows/`) and run one of the pre-built multi-agent examples to see graph-based orchestration in action.
03Launch the DevUI package (`python/packages/devui/`) against your sample agent to visually inspect message flow, debug agent handoffs, and identify where to insert human-in-the-loop checkpoints.

CODE INTELLIGENCE

bash

pip install agent-framework --pre
# .NET alternative:
# dotnet add package Microsoft.Agents.AI

FIELD OPERATIONS

Customer Support Triage Pipeline

Build a multi-agent workflow where a routing agent classifies incoming support tickets by category and urgency, then dispatches to specialized agents (billing, technical, returns) that draft responses — with a human-approval checkpoint before any response is sent to the customer.

Automated Research Briefing System

Create a graph-based workflow with a search agent, a summarization agent, and a formatting agent that runs on a schedule, pulling news and competitor data, summarizing findings, and delivering a structured daily briefing document to a Slack channel or email.

STRATEGIC APPLICATIONS

→A legal services firm uses a multi-agent pipeline where one agent extracts clauses from uploaded contracts, a second agent checks each clause against a compliance rulebook, and a third agent generates a risk summary report — cutting contract review time from hours to minutes.
→An e-commerce company orchestrates post-purchase workflows where agents handle order confirmation, inventory updates, shipping notifications, and customer follow-up emails as a coordinated graph, replacing a fragile chain of Zapier automations with a single auditable, checkpointable pipeline.