A local production-oriented architecture exercise for an AI incident-response system.
Alert / Incident → Planner → ExecutionPlan → Scheduler → Worker Registry → Parallel Workers → Evidence Aggregator → Hypothesis Generator → Investigation Assessment
If evidence is insufficient: → Planner investigates again
If evidence is sufficient: → Incident Decision → Risk Assessment → Policy Evaluation → Authorization → HITL / Automated Execution → Verification → Bounded Retry / Investigation Loop
Python 3.11+
python -m venv .venv