PulseAugur
中
实时 02:00:17
English(EN) LEANGUARD: Lean-Verified Guardrails for Tool-Using Agents

LeanGuard 为 AI 代理强制执行已验证的策略，减少提示注入

作者 PulseAugur 编辑部 · [2 个来源] ·

研究人员开发了 LeanGuard，一个新颖的运行时系统，旨在为使用工具的 AI 代理强制执行策略。该系统将工具调用与用 Lean4 编写的策略进行验证，防止代理执行被禁止的操作。LeanGuard 不仅阻止未经授权的调用，还为被拒绝的操作提供基于证明的修复，显著降低提示注入的成功率，同时不影响良性任务的性能。在 Qwen 模型上的测试表明，LeanGuard 能有效消除策略违规，并且其反馈机制可以提高任务成功率。 AI

影响 这项研究通过提供强大的策略执行和错误纠正机制，有望实现更安全、更可靠的 AI 代理。

排序理由 该集群描述了一篇详细介绍 AI 代理新系统的研究论文。

在 Hugging Face Daily Papers 阅读 →

AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →

LeanGuard 为 AI 代理强制执行已验证的策略，减少提示注入

本文如何被排名

Signal score
0 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Research
该集群描述了一篇详细介绍 AI 代理新系统的研究论文。
Source corroboration
2 independent sources
Multiple independent publishers reporting the same story raises confidence that it's real and newsworthy.
Topics
paper, safety, product
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
9 days old
Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.

完整方法见我们的编辑标准。

报道来源 [2]

  1. arXiv cs.CL TIER_1 English(EN) · 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… ·

    LEANGUARD：用于工具使用代理的精益验证保护栏

    arXiv:2611.02332v1 Announce Type: cross Abstract: Tool-using language-model agents often execute actions that their domain policies forbid. We present LeanGuard, a runtime that checks every proposed tool call against a policy written in Lean4 before the call reaches the environme…

  2. Hugging Face Daily Papers TIER_1 English(EN) ·

    LEANGUARD：面向使用工具的智能体的精益验证防护栏

    LeanGuard is a verified runtime that blocks policy-violating tool calls by language-model agents and can repair them via proof-guided search, reducing unauthorized executions and prompt-injection success without harming benign task performance.

相关实体

相关话题