Skip to content

Translation in progress Some entries still include text in the original language.

Oct 13 · Tuesday

All papers

Found 23 papers

Another 216papers do not have category tags yet and are not included in filtered results.

  1. DefensesarXiv 2610.07657

    DEFER1 多智能体防御研究报告受控测试中的攻击成功率下降

    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
    Evaluated models
    Qwen3-235B-A22B-Instruct-2507, gpt-oss-120b, Llama-4-Scout-17B-16E-Instruct and others5
    AbstractDEFER 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。

    Paper analysisRead analysis
  2. DefensesarXiv 2610.02861

    论文提出在 Kubernetes 上保护 AI Agent 的七层纵深防御框架与参考架构

    提出 Kubernetes 运维智能体的七层纵深防御框架与参考架构

    Published
    Setting
    AI Agent
    Abstract把运维智能体当不可信租户,用七层控制在完全攻破假设下约束行动。

    把 Kubernetes 上的运维智能体当作不可信租户,在假定提示注入已完全控制其工具选择的前提下,用基础设施重建数据面与控制面的边界。要解决的是:智能体读日志、注解、工单或 MCP 工具描述时,这些内容可以变成它带着自身凭据执行的命令。模型对齐和分类器只降低概率。

    Paper analysisRead analysis
  3. OtherarXiv 2609.12413

    SoK 综述:智能体 AI 时代的越狱攻击、防御与实践考量

    系统梳理智能体执行链路中的越狱攻击与防御并比较推理时防御

    Published
    Setting
    AI Agent
    Evaluated models
    Qwen3.5-4B, Llama-3.1-8B-Instruct
    AbstractSoK 将越狱重放到智能体流水线,并用两种模型检验回复层防御的边际与代价。

    Mia 等人把越狱安全从单轮回复扩展到智能体的规划、记忆、工具和智能体间通信,并问既有结论在更强原生对齐下是否还成立。

    Paper analysisRead analysis
  4. DefensesarXiv 2609.04875

    Forgetting without restart: execution-state forgetting for stateful LLM agents

    Proposes execution-state forgetting, using provenance-guided selective replay to purify stateful agents

    Published
    Setting
    AI Agent
    Evaluated models
    Llama-3.1-8B-Instruct, Qwen2.5-7B-Instruct, Mistral-7B-Instruct-v0.3
    AbstractSelective replay concatenates the clean prefix with the sanitized suffix, and main-table metrics match full reset.

    This work defines forgetting in long-horizon agents as runtime counterfactual equivalence, rather than deleting a plaintext memory.

    Paper analysisRead analysis
  5. ActBench: A Self-Evolving Benchmark for Behavioral Safety of Collaborative Agents

    构建自演化基准 ActBench,评测协作智能体的行为安全

    Published
    Setting
    AI Agent
    Evaluated models
    Claude-Opus-4.8, Claude-Sonnet-4.6, GPT-5.5 and others15

    AbstractActBench 通过自演化搜索与双证据重构评测轨迹行为安全,揭示基础模型决策差异主导了协同智能体的操作风险。

    Paper analysisRead analysis
  6. AttacksarXiv 2607.06595

    When Agents Remember Too Much: Memory Poisoning Attacks against Large Language Model Agents

    Proposed GhostWriter memory poisoning, which writes into the agent's long-term memory via email and calendar and activates across sessions

    Published
    Setting
    AI Agent
    Evaluated models
    GPT-5.4-mini, DeepSeek-V4-Flash, Gemini-2.5-Flash and others4
    Paper analysisRead analysis
  7. AttacksarXiv 2607.05189

    Researchers propose MemGhost, achieving covert memory injection into persistent personal agents

    提出 MEMGHOST,用单封邮件对持久个人 Agent 做跨会话隐蔽记忆注入

    Published
    Setting
    AI Agent
    Evaluated models
    GPT-5.4, Claude Sonnet 4.6, Claude Sonnet 4.5 and others6
    Abstract总结了论文的研究背景、双代理优化方法 MEMGHOST、全周期实验结果以及作者对记忆信任边界的启示。

    本文研究了持久化个人智能体面临的隐蔽记忆注入威胁,提出了一套兼顾写入、隐蔽与跨会话生效的自动化生成框架与全周期评测基准。

    Paper analysisRead analysis
  8. DefensesarXiv 2606.24322

    Protecting LLM agent long-term memory from poisoning: tamper-evident, source-bound permissions and machine-checkable guarantees

    提出 TMA-NM,以写时来源绑定抵御 Agent 长期记忆投毒

    Published
    Setting
    AI Agent
    Evaluated models
    gpt-5-chat, gpt-4o-mini, claude-opus-4.1 and others8
    AbstractTMA-NM 用来源绑定的行动权限挡住记忆洗白。

    TMA-NM 把 LLM 智能体长期记忆的行动权限绑到写时来源,用来挡住跨会话记忆投毒。

    Paper analysisRead analysis
  9. AttacksarXiv 2606.04329

    研究者系统研究 LLM 智能体记忆投毒攻击并提出 MPBench 基准

    系统研究 LLM 智能体记忆投毒并构建基准 MPBench

    Published
    Setting
    AI Agent
    Evaluated models
    Meta-Llama-3.1-70B-Instruct, Deberta-v3-base, Llama-3.1-8B-Instruct and others5

    Abstract系统研究了智能体记忆投毒风险,揭示了写入渠道漏洞,构建了 MPBench 基准并指出提示注入防御的局限。

    Paper analysisRead analysis
  10. AttacksarXiv 2605.15338

    Hidden in Memory: Latent Memory Poisoning of LLM Agents

    提出潜伏记忆投毒,诱导 Agent 把外部文档写成持久伪造记忆

    Published
    Setting
    AI Agent
    Evaluated models
    GPT-5.4, GPT-5.5, Claude Sonnet 4.6 and others12
    Abstract持久记忆成为跨会话攻击面:通用载荷可写入伪造记忆,并在相关后续任务中影响行为。

    sleeper memory poisoning 研究带持久记忆的 LLM 助手:外部文档能否让助手存下一则伪造的用户记忆,并在原文档消失后的新会话里被取回、用来改变回复或动作。

    Paper analysisRead analysis
  11. Trojan Hippo Bench: A Dynamic Benchmark for Persistent Memory Attacks and Defenses on LLM Agents

    构建 Trojan Hippo Bench,评测 Agent 持久记忆投毒与外泄

    Published
    Setting
    AI Agent
    Evaluated models
    Gemini 3.1 Pro, GPT-5-mini, GPT-5

    Abstract论文提出了评测智能体持久化记忆攻击与防御的 Trojan Hippo Bench,揭示了潜伏攻击威胁及内存防御的安全与效用权衡。

    Paper analysisRead analysis
  12. 论文提出 eTAMP 攻击:网页内容注入可跨会话污染 Web Agent 记忆

    提出 eTAMP,用网页内容跨会话投毒 Web Agent 轨迹记忆并诱导越权操作

    Published
    Setting
    web agent
    Evaluated models
    GPT-5-mini, GPT-5.2, GPT-OSS-120B and others6
    Abstract论文提出了基于环境注入的跨任务记忆投毒攻击 eTAMP,作者报告智能体在环境压力下易感性增加且能力与安全并不协同。

    该论文评估了基于大语言模型的 Web 智能体在利用历史交互轨迹进行记忆增强时的安全脆弱性。

    Paper analysisRead analysis
  13. DefensesarXiv 2603.12644

    Revealing security threats and defense architectures for autonomous agents: an OpenClaw case study

    Reviews threats to autonomous agents and proposes a full-lifecycle defense blueprint, FASA

    Published
    Setting
    AI Agent
    AbstractUsing OpenClaw as an example, it divides risks into three layers and uses FASA to describe defense from input to system execution.

    This is a threat review and defense architecture report using OpenClaw as a case study, not a scored attack or defense evaluation.

    Paper analysisRead analysis
  14. DefensesarXiv 2603.11619

    Taming OpenClaw: Security Analysis and Mitigation of Threats in Autonomous LLM Agents

    Proposes a defense-in-depth aligned with the five stages to mitigate cross-stage threats to autonomous agents

    Published
    Setting
    AI Agent
    AbstractThis is a lifecycle threat analysis and layered mitigation design for OpenClaw, with the cases serving as qualitative evidence.

    This is a lifecycle threat analysis of OpenClaw, accompanied by a set of layered mitigation designs. The problem addressed is: when autonomous agents connect instant messaging, persistent memory, third-party plugins, and high-privilege execution, attacks are no longer limited to single-turn jailbreaks but can compound across multiple stages from initialization to execution. The authors argue that defenses hardening only individual interfaces cannot stop such cross-temporal attacks.

    Paper analysisRead analysis