"Agentjacking" is the latest demonstration of how easily attackers can exploit an AI agent's inability to differentiate between content and instructions.

This is an ingested summary. Full reporting is available at the primary source below.