<<<<<<< HEAD
SafeAgentSOC is a GRC-aware, tool-constrained AI SOC co-analyst prototype for reducing SOC alert fatigue through SIEM telemetry, alert normalization, incident case-building, context enrichment, LLM-assisted hypothesis generation, graph validation, risk scoring, GRC policy guardrails, constrained action recommendations, and human-in-the-loop approval.
Phase 2 started: Scenario Design, Emulation, and Dataset Creation.
Sprint 0: Workspace, dataset governance, and safety setup.
Create a controlled, labeled, MITRE-mapped Wazuh alert dataset across benign baseline, manual adversary emulation, Atomic Red Team validation, MITRE Caldera campaigns, and simulated-only high-risk gaps. The dataset will support later alert normalization, clustering, context enrichment, MITRE mapping, LLM hypothesis evaluation, graph validation, and policy-safe response evaluation.
=======
f8467cef03b5888f73ff6246979e76d32f73b841