LeanGuard用Lean验证智能体工具调用
Sreehari Sankar及19位共同作者提出LeanGuard,这一运行时会在工具调用进入环境前,依据Lean4策略检查每次调用,并用状态谓词和限定作用域的过去时间算子表达约束。作者称其评估采用失败关闭机制,判定过程被证明与LeanLTL语义等价,证明引导修复还可在不调用模型的情况下寻找合规修改。在使用五个Qwen模型的Tau-Bench、DTap和电信任务实验中,论文报告LeanGuard未执行策略禁止的调用,将DTap直接提示注入成功率从57.9%降至5.9%,且未改变电信任务的正常成功率。
该论文提出了一种在工具型智能体行动进入环境前进行形式化约束的失败关闭方案,并给出多基准评估与证明引导修复机制。
Computer Science > Artificial Intelligence
arXiv:2611.02332 (cs)
[Submitted on 2 Nov 2026]
Title:LEANGUARD: Lean-Verified Guardrails for Tool-Using Agents
Authors:Sreehari Sankar, Debargha Ganguly, Vikash Singh, Mohsen Hariri, Weicong Chen, Yanyan Zhang, Shouren Wang, Wang Yang, Chaoda Song, Alexander Nemecek, Osama Zafar, Idan Davidovich, Yuting Shao, Kai Ye, Biyao Zhang, Xinpeng Li, Joe Stubbs, Erman Ayday, Jing Ma, Vipin Chaudhary
View a PDF of the paper titled LEANGUARD: Lean-Verified Guardrails for Tool-Using Agents, by Sreehari Sankar and 19 other authors