Early Detection of Distributed Backdoors in Multi-Agent LLM Systems: A Characterization Study

본문 미리보기

arXiv:2607.24893v1 Announce Type: new Abstract: Multi-agent LLM systems can be attacked by a payload that no single agent ever holds in full: a poisoned tool hides encrypted fragments in its observations, spreads them across several agents, and an external step reassembles and executes them after the run. Per-step safety checks that judge each action in isolation may fail to recognize the complete distributed payload. We investigate how early such an attack can be detected while the run is stil