After 30+ years in critical infrastructure security (power grid, banking), I got frustrated with SIEM tools that generate mountains of false positives. So I built something different.
ASAF uses information geometry — specifically Fisher Information Matrices and geodesic analysis on statistical manifolds — to model network behavior. Instead of pattern matching against known signatures, it measures how the "shape" of network traffic deviates from its natural geometry.

Key technical details:

C++17/Eigen3, runs air-gapped (no cloud dependency)
Ihara zeta function for network topology analysis
Geodesic extrapolation via RK4 for predictive threat detection
Maps findings to MITRE ATT&CK framework for actionable reports

The core idea: network traffic lives on a statistical manifold. Normal traffic follows geodesics. Attacks create curvature anomalies that are mathematically detectable before traditional IDS/IPS triggers.
Built for industrial/critical infrastructure environments where air-gap is mandatory.
Happy to discuss the math or architecture. Been running it on real infrastructure in Mexico.

Contact: entfernten.galaxie@gmail.com | https://consultoria.aivoix.mx