跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2509.14956 · v2

Sentinel Agents for Secure and Trustworthy Agentic AI in Multi-Agent Systems

首次提交 2025-09-18 21:39
版本更新 2026-11-06 01:05
雷达观测 2026-11-09 10:16
arXiv ↗PDF ↗

摘要

Multi-agent systems (MAS) built on large language models (LLMs) increasingly share a common conversational space, in which a single adversarial or erroneous message can propagate from one agent to many. This paper proposes Sentinel Agents, a dedicated security layer that monitors inter-agent communication with rule-based checks, behavioral analysis and LLM-based semantic inspection, and reports to a Coordinator Agent that enforces policy, for example by quarantining misbehaving agents. We describe the architecture, its deployment patterns and monitoring layers, and we evaluate a continuous-listener prototype on a balanced corpus of 1,000 synthetic messages (500 attacks across five families and 500 benign messages) with two local LLMs. LLM inspection with Gemma 4 12B reached F1 0.86 at a false-positive rate of 0.11, against 0.38 for rules alone. Combining components by logical OR did not improve on it, while a cascade in which only high-precision rules act on their own matched it. Calibrating the thresholds of the more conservative Qwen3 8B on a development split raised its held-out F1 from 0.56 to 0.80, and on multi-turn conversations the behavioral check detected patterns the rules missed but could not tell legitimate repetition from probing. The value of combining detectors therefore depends on the model, on its calibration and on the aggregation rule, and the results yield practical guidance for deploying Sentinel Agents.

作者备注:23 pages, 11 figures. v2: the preliminary evaluation of v1 is replaced by a new evaluation on a balanced 1,000-message corpus with ablation, local LLM benchmark, aggregation and calibration analyses; paper restructured; references updated

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

本站首次收录为 v2,暂无更早版本可对照。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v2 · 2026-11-06 01:05 · 最新观测版本

Sentinel Agents for Secure and Trustworthy Agentic AI in Multi-Agent Systems

Multi-agent systems (MAS) built on large language models (LLMs) increasingly share a common conversational space, in which a single adversarial or erroneous message can propagate from one agent to many. This paper proposes Sentinel Agents, a dedicated security layer that monitors inter-agent communication with rule-based checks, behavioral analysis and LLM-based semantic inspection, and reports to a Coordinator Agent that enforces policy, for example by quarantining misbehaving agents. We describe the architecture, its deployment patterns and monitoring layers, and we evaluate a continuous-listener prototype on a balanced corpus of 1,000 synthetic messages (500 attacks across five families and 500 benign messages) with two local LLMs. LLM inspection with Gemma 4 12B reached F1 0.86 at a false-positive rate of 0.11, against 0.38 for rules alone. Combining components by logical OR did not improve on it, while a cascade in which only high-precision rules act on their own matched it. Calibrating the thresholds of the more conservative Qwen3 8B on a development split raised its held-out F1 from 0.56 to 0.80, and on multi-turn conversations the behavioral check detected patterns the rules missed but could not tell legitimate repetition from probing. The value of combining detectors therefore depends on the model, on its calibration and on the aggregation rule, and the results yield practical guidance for deploying Sentinel Agents.

查看此版本解读 →查看此版本原文 ↗