nasim

A PPO agent leveraging reinforcement learning performs Penetration Testing in a simulated computer network environment. The agent is trained to scan for vulnerabilities in the network and exploit them to gain access to various network resources.

Neuro-symbolic RL agent that learns to pentest networks it has never seen — GPT-4o compiles CVE preconditions into a Z3 action mask over a GraphSAGE PPO policy. Zero-shot attack-graph transfer, negatives disclosed.

All 3 Type camera , Mobile and TV create by PHP