microsoft/CyberBattleSim

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1784 stars · 285 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

CyberBattleSim is a Python-based experimentation and research platform from Microsoft that simulates abstract enterprise network environments where automated attacker and defender agents interact. It provides an Open AI Gym interface so reinforcement learning algorithms can train agents to exploit vulnerabilities, move laterally across networks, and defend against such attacks.

Use cases

• train reinforcement learning agents for cyber attack simulation
• research lateral movement attacks in simulated networks
• benchmark RL algorithms on cybersecurity environments
• simulate attacker versus defender interactions in enterprise networks
• study how network topology affects cyber attack propagation
• build a custom gym environment for security research

When to choose

• you need a safe, abstract sandbox for cybersecurity RL research
• you want to compare reinforcement learning agents on attack/defense tasks
• you are studying the effect of network topology and vulnerabilities on lateral movement

When to avoid

• you need realistic network traffic or direct application to real-world systems
• you want production-grade penetration testing tools
• you need a polished game or commercial training product

Facets

library · maturity active

simulation reinforcement-learning agent-framework security penetration-testing security artificial-intelligence reinforcement-learning simulation python cross-platform cybersecurity-simulation openai-gym attack-defense lateral-movement network-simulation research-platform research

Member repositories

For agents

markdown · JSON · MCP: product_card(name="microsoft/CyberBattleSim")