Skip to content

notebooks

Directory actions

More options

Directory actions

More options

Latest commit

 

History

History

notebooks

README.md

Agent OS Notebooks

Interactive tutorials for learning Agent OS step-by-step.

🚀 Getting Started

pip install agent-os-kernel[full] jupyter
jupyter notebook

📚 Notebook Index

# Notebook Description Time Prerequisites
01 Hello Agent OS Your first governed agent 5 min None
02 Episodic Memory Persistent agent memory 15 min 01
03 Time-Travel Debugging Replay and debug decisions 20 min 01
04 Verification Detect hallucinations with CMVK 15 min 01
05 Multi-Agent Coordination Trust between agents (IATP) 20 min 01
06 Policy Engine Deep dive into policies 15 min 01

🎯 Learning Paths

Path 1: Quick Start (30 min)

For developers who want to get up and running fast:

  1. 01 - Hello Agent OS
  2. 06 - Policy Engine

Path 2: Agent Memory & Debugging (50 min)

For developers building agents that learn:

  1. 01 - Hello Agent OS
  2. 02 - Episodic Memory
  3. 03 - Time-Travel Debugging

Path 3: Multi-Agent Systems (55 min)

For developers building complex agent systems:

  1. 01 - Hello Agent OS
  2. 04 - Verification
  3. 05 - Multi-Agent Coordination

Path 4: Complete Course (1.5 hours)

All notebooks in order for comprehensive understanding.

📦 Installation by Notebook

Each notebook lists its specific dependencies, but here's a quick reference:

# Minimal (notebooks 01, 06)
pip install agent-os-kernel

# Episodic Memory (notebook 02)
pip install agent-os-kernel emk

# Verification (notebook 04)
pip install agent-os-kernel[cmvk]

# Multi-Agent Coordination (notebook 05)
pip install agent-os-kernel[iatp]

# Everything
pip install agent-os-kernel[full]

🔧 Running Notebooks

Option 1: Jupyter Notebook

jupyter notebook

Option 2: JupyterLab

pip install jupyterlab
jupyter lab

Option 3: VS Code

  1. Install the Jupyter extension
  2. Open any .ipynb file

Option 4: Google Colab

Upload notebooks to Google Colab for cloud execution.

📖 What You'll Learn

Core Concepts

  • Kernel Space vs User Space: How Agent OS enforces policies
  • Policy Engine: Defining what agents can and cannot do
  • Signals: SIGKILL, SIGSTOP, SIGCONT for agent control

Memory Systems

  • Episodic Memory (EMK): Immutable record of agent experiences
  • Memory Compression: Sleep cycles to distill knowledge
  • Negative Memory: Tracking failures to avoid repeating mistakes

Verification & Trust

  • CMVK: Verification for hallucination detection
  • IATP: Cryptographic signing for multi-agent trust
  • Consensus: Multi-model agreement protocols

Debugging

  • Flight Recorder: Capture every decision point
  • Time-Travel: Replay agent state at any moment
  • Audit Trails: Complete logging of policy decisions

💡 Tips

  1. Run cells in order: Each notebook builds on previous cells
  2. Read the output: Explanations are in the output, not just the code
  3. Experiment: Modify the code and re-run to see what happens
  4. Check the docs: Each notebook links to detailed documentation

🤝 Contributing

Found an issue or want to add a notebook?

📚 Additional Resources