Agents that passively watch a system and use LLM reasoning to flag anomalies and raise alerts. - View it on GitHub