Event-Driven Temporal Graph Networks for Asynchronous Multi-Agent Cyber Defense in NetForge_RL

NetForge_RL is currently a nascent research-oriented project (0 stars, 7 days old) focused on a highly specialized niche: bridging the gap between simulated cyber wargaming and real-world SOC operations. Its primary contribution is the 'Event-Driven Temporal Graph Network' (ED-TGN) approach, which moves beyond the synchronous 'turn-based' abstraction common in legacy simulators like Gym-CybORG or CALDERA. While the technical approach is sophisticated, it lacks the ecosystem or community adoption (stars/forks) to be considered defensible. The moat resides in the fidelity of the 'NetForge_RL' simulator itself; if this simulator becomes the gold standard for high-fidelity cyber-telemetry simulation, its defensibility would rise. Currently, it is a reference implementation for an academic paper. Frontier labs (OpenAI/Anthropic) are unlikely to compete here as it requires deep domain expertise in network protocol physics and SOC operations, which is too niche for general-purpose model providers. The displacement risk comes from other academic groups or specialized cyber-AI startups (e.g., those working on autonomous blue-teaming) who might release more mature, integrated simulation platforms.