// about
About this room
This module explores how autonomous AI agents become unwitting accomplices in cyberattacks through tool abuse. You will learn how the Confused Deputy problem manifests in agentic architectures, allowing attackers to manipulate tool parameters, chain authorized actions, and escalate privileges. Through advanced red teaming scenarios, you will discover how to exploit and defend against these critical execution-layer vulnerabilities.
Room tasks
2 tasksEach task’s Rep is set individually — advanced tasks average +50. The +0 room bonus unlocks at 100%.