What to log when an AI agent acts on someone's behalf
Most agent stacks already log plenty: prompts, completions, tool calls, latency, token counts. That's useful for debugging. After an incident, though, a security reviewer will ask something the traces can't answer: was this action allowed, for this agent, on behalf of this person, against this record, at this time? If an agent can change things in systems that matter, each decision should record roughly this: The person who delegated or started the action. Which agent acted, down to the client o
🎯 Vulnerability: Web Security Vulnerability
💡 Defense Fix: Implement strict input validation, contextual output encoding, and enforce least-privilege authorization controls.