Paper Detail
Davis Brown, Samarth Bhargav, Arav Santhanam, Kasper Hong, Ivan Zhang, Matan Shtepel, Steffi Chern, Alexander Robey, Eric Wong, Hamed Hassani
Language models can find thousands of severe software vulnerabilities, and agents are increasingly being misused for cyberattacks. To avoid detection, attackers frequently distribute their misuse, splitting a harmful task across many user accounts so each individual transcript looks benign. Because safety monitors score only one agent context at a time, they are structurally blind to misuse that is only visible in aggregate, across many accounts. We show this gap is real by building, to our knowledge, the first distributed agent attack, a multi-agent scaffold that completes hard cybersecurity tasks while hiding the harmful objective across subagents with limited contexts, evading a standard monitor that catches it only a fifth as often as prior agent attacks. Towards a defense, we develop an online stateful monitor that uses real-time clustering to collect weak suspiciousness signals across many agent transcripts, and escalates only rarely to a language model that flags misuse across user accounts. In evaluations with large-scale simulated datacenter traffic, our monitor Pareto dominates standard monitors, catching distributed attacks 30% earlier and flagging cyber misuse before it reaches the most harmful stages. Crucially, this comes at negligible additional latency for ~99% of user traffic. This detection advantage persists but narrows as the benign background traffic grows very large. After an extensive red-teaming exercise, we improve the defense and surprisingly also find that it catches standard jailbreaks, since adaptive attackers reuse attack variants across accounts. Our results point toward a new class of safety monitors which reason over groups of users rather than isolated transcripts.
No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.
No ranking explanation is available yet.
No tags.
@article{brown2026stateful,
title = {Stateful Online Monitoring Catches Distributed Agent Attacks},
author = {Davis Brown and Samarth Bhargav and Arav Santhanam and Kasper Hong and Ivan Zhang and Matan Shtepel and Steffi Chern and Alexander Robey and Eric Wong and Hamed Hassani},
year = {2026},
abstract = {Language models can find thousands of severe software vulnerabilities, and agents are increasingly being misused for cyberattacks. To avoid detection, attackers frequently distribute their misuse, splitting a harmful task across many user accounts so each individual transcript looks benign. Because safety monitors score only one agent context at a time, they are structurally blind to misuse that is only visible in aggregate, across many accounts. We show this gap is real by building, to our know},
url = {https://arxiv.org/abs/2605.31593},
keywords = {cs.CR, cs.AI},
eprint = {2605.31593},
archiveprefix = {arXiv},
}
{}