Translation in progress Some entries still include text in the original language.
Oct 12 · Monday
Agent safety · Papers Found 9 papers “
About this collection Collection in progress We are building a large AI safety paper collection to help researchers find work related to their interests. Papers are collected automatically andfiltered only for relevance to AI safety , without judging quality, novelty or reading value. The default “Annotated first” shows papers with tags or published paper analyses first. Each group is ordered by original publication date, newest first, with undated papers last. “Newest first” uses publication dates only. This order is not a quality ranking.
One-line summaries, tags and paper analyses are prepared by AI from the papers. Language models may have systematic biases and produce hallucinations.We plan to carry out human annotation and review to reduce errors in this content.
提出义务识别基准 ObligationBench，并训练护栏识别 Agent 未履行的安全义务
Published Oct 8
Evaluated models ObligationGuard , Claude-Opus-4.8 , GPT-5.6-Sol and others17 Abstract 义务是护栏缺口。作者认为编码智能体的安全不能只靠拦截禁止动作，还要找出终止时仍未做的安全关键动作，并将其称为义务。
实现 DEFER，以确定性检查优先拦截多智能体流水线的内容攻击
Published Oct 6
Evaluated models Qwen3-235B-A22B-Instruct-2507 , gpt-oss-120b , Llama-4-Scout-17B-16E-Instruct and others5 Abstract DEFER 把确定性检查放在 judge 之前。DEFER 用来测量多智能体授权流水线里，确定性规则在何处结束、LLM judge 从何处开始。
提出 ASA 与 AttestMCP，度量并降低 LLM 对输入序列扰动的敏感
Published Oct 2
Evaluated models Gemma-3-27B-Instruct , Gemma-3-4B-Instruct , Llama-3.2-3B-Instruct and others12 对比类型化评审与生成式评审对智能体轨迹的安全判定
Published Sep 28
Evaluated models JEV , GLM-5.2-Tencent , DeepSeek-V4-Flash and others5 Abstract JEV 的四基准平均正类 F1 最高，延迟和估计费用更低，但分数据集与精确率有取舍。在已保存的工具使用智能体轨迹上，比较类型化评审 JEV 与四种生成式评审的回溯安全分类。
构建支付欺诈基准 ACB，评测花钱智能体的欺诈检测
Published Sep 27
Evaluated models gordonguard , pattern judge Abstract ACB 把支付欺诈与智能体操纵分开测量，作者强调观察权限、误报预算和微支付复核成本。ACB 是一套智能体支付欺诈基准与检测参考栈 ，目标是衡量钱是否被不当提取，而不仅是智能体是否接受了恶意指令。
提出 ATF 与 NTR，过滤或恢复工具集以阻止 Agent 调用攻击工具
Published Sep 14
Evaluated models Gemma2-9B , Qwen2-7B , LLaMA3-8B and others7 评测 Agent 记忆系统的软撤回执行情况，并提出陈旧检索防护
Published Sep 8
Evaluated models GPT-5.5 , Grok-4 , DeepSeek-V4-Flash and others9 Abstract 系统揭示了记忆系统软撤回在检索时普遍未执行的安全漏洞，证明写回会击穿现有防御，并提出读取端防护作为必要控制手段。
Proposes a static-first validator, CARE, that decides before dispatch whether an agent's shell command may be executed
Published Jul 22 Proposes out-of-process dual-channel comparison classification to intercept erroneous clicks by computer-use agents
Published Mar 16
Evaluated models multi-modal-embed-small Abstract Perception errors can be exploited by swapping screen pixels.The authors frame CUA's perception failure as a security vulnerability, visual confused deputy : the agent authorizes click(x,y) based on a misread screen, and the actual landing point differs from the element it believes it is clicking.
You have reached the end