Bounded Agents: Delegation Security for Multi-Agent AI Systems
AI · Security · 2026-08-21 00:32:00
The paper introduces "bounded agents" to enhance security in multi-agent AI systems by preventing malicious actions and authority misuse during task delegation.