C CORRIGIBILITY Agent Incident Response Lab
AGENT INCIDENT RESPONSE

When an agent goes wrong,
can you contain and recover?

Simulate a fictional agent incident and inspect whether your control plane can detect, contain, revoke, roll back, and safely restore operation.

DETECT→CONTAIN→REVOKE→ROLL BACK→RECOVER→VERIFY
CONTAINMENT READINESS
78 /100

Strong containment

SYNTHETIC BLAST RADIUS
34
Contained Residual impact index

03 · RESPONSE TIMELINE

Incident → verified recovery

Estimated recovery: 28 min
04 · RESPONSE SEQUENCE

Recommended containment path

Ordered controls
RECOVERY READINESS 74

EVIDENCE QUALITY 84

ROLLBACK FEASIBILITY 71

PRIMARY RESPONSE GAP

POST-INCIDENT RULE

05 · INCIDENT RECORD

Machine-readable JSON


        


        

          
            
              How the Incident Response Lab works
              Synthetic recovery heuristics, not a security certification
            
            ＋
          
          

            

              The simulator combines incident severity, propagation and external side effects
              with detection, revocation, rollback, checkpointing, logging and human response
              capabilities to generate deterministic response signals.
            

            

              Real incident plans should additionally cover system-specific authentication,
              legal obligations, data retention, communications, backups, threat models,
              forensic procedures and independent security review.
            

          

        

      

    


    
  


  
Copied