A syllabus for training AI to attack and defend networks

What it does

This is a curated awesome-list that catalogs research and tools at the intersection of reinforcement learning and cybersecurity. It gathers RL-specific environments—such as CSLE, C-CyberBattleSim, MininetGym, and PenGym—alongside papers, books, and talks. The maintainer filters out general ML work, so every entry explicitly uses RL for attack, defense, or network control.

The interesting bit

The narrow gatekeeping is what makes it useful: general machine-learning-for-security lists already exist, but this one insists on RL-only. That turns it into a quick index for researchers who need Gymnasium-compatible cyber ranges or game-theoretic frameworks rather than another malware-classification survey.

Key highlights

• Catalogs specialized RL environments such as CSLE, C-CyberBattleSim, MininetGym, CyGym, and PenGym, many with linked papers and documentation.
• Organized into environments, papers, books, blog posts, talks, and miscellaneous resources.
• Explicitly excludes general ML methods; the scope is strictly reinforcement learning applied to security.
• Many entries include direct links to arXiv preprints, IEEE papers, theses, and ReadTheDocs sites.

Verdict

Researchers building or benchmarking autonomous cyber-defense agents will find the simulator inventory alone worth bookmarking. If you are looking for general ML security datasets or production-grade SOAR playbooks, this is not your list.