Skip to content

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

Oct 12 · Monday

Agent safety · Papers

Found 4 papers
Filter2

Attack type

Prompt injection4Poisoning and backdoors4

Defense type

Permissions and approval3

Impact

Unauthorized actions4

System type

AI Agent4

Attack surface / entry point

User input4Web pages and interfaces3Documents and messages3Tool responses4Memory writes9Distribution channels3

Layer

Application layer4

System component

MCP and skills7Memory4Guardrails and judges3Permissions and sandboxing22
Select any value within a category; different categories must all match. Counts include the other filters you have selected.
Attack surface / entry point：Tool responsesSystem component：MemoryTopic: Agent safetyClear all
  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

    arXiv
    2610.07657
    Published
    Layer
    Application layer
    Setting
    Multi-agent systems
    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

    arXiv
    2610.02861
    Published
    Layer
    Application layer
    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. Evaluation and benchmarksarXiv 2608.09476

    ActBench：面向协作智能体行为安全的自演化基准

    构建自演化基准 ActBench，评测协作智能体的行为安全

    arXiv
    2608.09476
    Published
    Layer
    Application layer
    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
  4. 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

    arXiv
    2606.24322
    Published
    Layer
    Application layer
    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
You have reached the end