← Back to Labs
Autonomous AI Agent Safety & Policy Controls
Trace the ReAct execution loop, indirect prompt injection attacks, tool privilege escalation, and Cedar policy enforcement
STEP 1 OF 6
ReAct Agent Context Initialization
The AI agent initializes its context window with system instructions, security guidelines, user prompt, and available tool schemas (such as file reading and exfiltration API endpoints).
Arrow keys to navigate · R to reset
Tap dots to jump to any step