Confused Deputy Attacks in AI Agent Systems
AI agents mistake untrusted content for instructions when they hold authority the author lacks.
Bianca Espinoza
Section
AI Agent Security
6 stories in AI Agent Security.
Code Execution Risks in LLM Tool Use
Securing AI agents requires isolation between instructions and data before models gain tool access.
Selah Beck
Human-in-the-Loop Controls for High-Risk Agent Actions
Approval checkpoints alone won't stop rogue AI agents from causing real-world harm.
Selah Beck
Tool Poisoning Attacks Against LLM Agents
Attackers exploit how AI agents trust tool descriptions without checking if they change mid-session.
Selah Beck
Least Privilege Principles Applied to AI Agents
Static roles can't handle how AI agents actually behave in practice.
Bianca Espinoza
Agent Memory and Context Persistence Security Risks
Persistent agent memory creates attack surfaces that outlast individual sessions.
Hugo Dvorak