四智能体协同守护多智能体系统安全，让规则可机器验证。
QuadSentinel: Sequent Safety for Machine-Checkable Control in Multi-agent Systems
- 用四类智能体将安全策略转为可观测状态的可验证规则。
- 在ST-WebAgentBench和AgentHarm上准确率提升，误报率下降。
- 无需改核心模型，适合快速部署到现有系统中。
基于大语言模型的智能体在执行复杂任务时面临安全风险，而人工编写的自然语言安全策略模糊且依赖上下文，难以转化为机器可验证的规则，导致运行时保护不可靠。本文提出 extsc{QuadSentinel}，一种由四个智能体（状态追踪器、策略验证器、威胁监视器、裁判）组成的协同机制，将安全策略表达为可验证的逻辑序列，并基于可观测状态构建机器可检查规则，在线实时执行。通过裁判逻辑与高效的 top-$k$ 谓词更新机制，优先处理关键检查并分层解决冲突，显著降低开销。在 ST-WebAgentBench（ICML CUA '25）和 AgentHarm（ICLR '25）基准测试中， extsc{QuadSentinel} 提升了守卫准确率与规则召回率，同时减少误报。相比 ShieldAgent（ICML '25）等单智能体基线，其整体安全控制效果更优。未来可无需修改核心智能体，仅通过分离且可机器验证的安全策略实现快速部署。代码将在 https://github.com/yyiliu/QuadSentinel 公开。
原文摘要 · Abstract (English)
Safety risks arise as large language model-based agents solve complex tasks with tools, multi-step plans, and inter-agent messages. However, deployer-written policies in natural language are ambiguous and context dependent, so they map poorly to machine-checkable rules, and runtime enforcement is unreliable. Expressing safety policies as sequents, we propose \textsc{QuadSentinel}, a four-agent guard (state tracker, policy verifier, threat watcher, and referee) that compiles these policies into machine-checkable rules built from predicates over observable state and enforces them online. Referee logic plus an efficient top-$k$ predicate updater keeps costs low by prioritizing checks and resolving conflicts hierarchically. Measured on ST-WebAgentBench (ICML CUA~'25) and AgentHarm (ICLR~'25), \textsc{QuadSentinel} improves guardrail accuracy and rule recall while reducing false positives. Against single-agent baselines such as ShieldAgent (ICML~'25), it yields better overall safety control. Near-term deployments can adopt this pattern without modifying core agents by keeping policies separate and machine-checkable. Our code will be made publicly available at https://github.com/yyiliu/QuadSentinel.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。