microsoft/CyberBattleSim
Microsoft's open-source platform for training attacker and defender agents using reinforcement learning in simulated enterprise network environments.
Not currently ranked — collecting fresh signals.
star history
CyberBattleSim is a research experimentation platform that simulates enterprise network environments where automated agents interact. The platform uses Open AI Gym interfaces to enable training of reinforcement learning agents, specifically attacker agents that exploit vulnerabilities to move laterally through a network and defender agents that detect and mitigate attacks by re-imaging compromised nodes. The simulation abstracts network topology and cyber security concepts to study multi-agent interactions and RL algorithm performance.
Frequently asked
- What is microsoft/CyberBattleSim?
- Microsoft's open-source platform for training attacker and defender agents using reinforcement learning in simulated enterprise network environments.
- Is CyberBattleSim open source?
- Yes — microsoft/CyberBattleSim is open source, released under the MIT license.
- What language is CyberBattleSim written in?
- microsoft/CyberBattleSim is primarily written in Jupyter Notebook.
- How popular is CyberBattleSim?
- microsoft/CyberBattleSim has 1.8k stars on GitHub.
- Where can I find CyberBattleSim?
- microsoft/CyberBattleSim is on GitHub at https://github.com/microsoft/CyberBattleSim.