{"name":"sentinelops","description":"SentinelOps is a multi-step agentic reinforcement-learning environment that simulates an AI security analyst monitoring surveillance camera feeds. The agent must inspect frames, navigate temporal sequences, switch between cameras, detect anomalies, classify threat severity, and decide whether to escalate incidents or dismiss false alarms.","version":"1.0.0","author":"Team Adaptrix","license":"MIT","tags":["surveillance","incident-response","multi-step","agentic","reinforcement-learning","security"]}