Morning Edition · Tuesday, June 30, 2026Published at 6:59 AM EDT · New York

Paper Shows Tool Use Lets Multi-Agent LLMs Collude Through Undetectable Steganography

Monitoring plain-text communication between agents may not catch secret coordination once those agents can call tools.

A new paper, Tool Use Enables Undetectable Steganography in Multi-Agent LLM Systems, targets a defense that many agent-safety designs rely on: monitoring the plain-text messages that agents exchange. The authors argue that once agents can c…