AutoGen Magentic-One Sample with Security best practices.

This sample demonstrates how to use the AutoGen Magentic-One with security best practices like:

• Code execution in Docker: Making sure the AutoGen m1 can only execute code in a Docker container.
• Safeguard File Access: Making sure the AutoGen m1 can only access the file from a safe directory, in this sample it's the current directory. It can be changed by modifying the logic of safe directory in _secured_markdown_file_browser.py file.
• Human Oversight for each critical action: Making sure the AutoGen m1 can only execute code or open a site or access a file after human approval. It can be changed by modifying the description of human agents in _secured_m1.py file and in the next agent selection prompt in _secured_m1_orchestrator.py file.
• Termination: Making sure the AutoGen m1 can be terminated by human agents at any time. It can be changed by modifying the termination command in _secured_m1.py file.

For more information on how to use this sample, please refer:

• How to Use Magentic One

Prerequisites

• Python 3.12 or later
• Docker Desktop
• Azure OpenAI Service

How to run

1. Clone the repository
2. Install the dependencies

   pip install -r requirements.txt

3. Create a .env file in the root directory by copying the .env.example file and filling in the required values.
4. Run some of the example with security best practices and see the result

   python secured_m1.py --secure "Find the best restaurant in town"
   python secured_m1.py --secure "Get Microsoft Stock Trend for last 5 years and plot the graph"
   python secured_m1.py --secure "Read the TEST.CSV file from the parent directory and summarize the content"