NetForge_RL bridges the simulation-to-reality gap in AI-powered cybersecurity by enabling multi-agent reinforcement learning systems to train in mock environments and deploy directly against live security threats.

Introduces NetForge_RL, a high-fidelity cyber operations simulator designed to overcome the Sim2Real gap that has blocked MARL policies from moving from simulated wargames to real Security Operations Centers (SOCs)

Reformulates network defense as an asynchronous, continuous-time Partially Observable Semi-Markov Decision Process (POSMDP) with realistic network protocol physics and noisy telemetry instead of clean state vectors

Features a dual-mode engine supporting both high-throughput MARL training in a mock hypervisor and zero-shot evaluation against live exploits in a Docker hypervisor

Enforces Zero-Trust Network Access (ZTNA) constraints and requires defenders to process NLP-encoded SIEM telemetry for more authentic operational scenarios