Three agents. One mission. Zero tolerance.
Three autonomous agents. Each with a distinct mandate. Together, they form an unbroken chain of oversight — from first detection to final enforcement.
Polls the entire AI Shield module stack every 30 seconds. Monitors every AI agent in the environment for behavioural drift, anomalies, and policy violations. First responder. First line of defence.
Triggered by RSSA-1 escalations. Conducts deep investigations across the full module stack — tracing attack chains, correlating signals, classifying threats. The Anthropic distillation attack involved 16 million exchanges across 24,000 fraudulent accounts. That's the scale RSSA-2 was built for.
Oversight of the entire RSSA force and all agents in the environment. Maintains fleet threat posture. Sole authority to trigger Module 99. Every enforcement decision cryptographically logged. Can terminate a compromised RSSA-1 or RSSA-2 if needed.
RSSA is wrapped and secured by AI Shield — monitored by the same 101 modules it uses to monitor others. No single agent has unchecked authority. The chain of custody is complete.
Every AI agent your organisation deploys is a potential attack vector. RSSA monitors all of them. Simultaneously. At machine speed.