Skip to content

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

Oct 13 · Tuesday

All papers

Found 12 papers
Filter4

Attack type

Prompt injection5Poisoning and backdoors5

Defense type

Model hardening3

Impact

Unauthorized actions5

System type

Models and APIs5AI Agent8coding agent7

Attack surface / entry point

User input65Documents and messages4Tool responses3Tool and skill descriptions5Code repositories and dependencies7Memory writes3Images, audio and video17Training data5Weights and gradients3Distribution channels6

Layer

Application layer7Training and data layer5
Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 227papers have no tags yet and are not included in filtered results.
System type：Models and APIsSystem type：coding agentAttack surface / entry point：Code repositories and dependenciesAttack surface / entry point：Training dataClear all

Another 227papers do not have category tags yet and are not included in filtered results.

  2. Evaluation and benchmarksarXiv 2609.36817

    pikit：面向间接提示注入研究与评测的可组合工具包

    构建可组合工具包 pikit，评测编码智能体的间接提示注入

    arXiv
    2609.36817
    Published
    Layer
    Application layer
    Setting
    coding agent
    Abstractpikit 提供组合式间接注入评测。

    pikit 是面向间接提示注入研究的组合式工具包，主要贡献是组织攻击、投递通道、防御和判定流程，而非提出新的攻击优化算法。

    Paper analysisRead analysis
  4. Evaluation and benchmarksarXiv 2609.32616

    Cave-Bench：虚假指控下 LLM 智能体最高 60.06% 的运行会破坏正确工作

    提出 CAVE-BENCH，评测智能体在虚假指责下破坏已完成的正确工作

    arXiv
    2609.32616
    Published
    Layer
    Application layer
    Setting
    coding agent
    Evaluated models
    Claude-Sonnet-5, Claude-Opus-5, GPT-5.6-Sol and others14

    Abstract论文提出了评估智能体在虚假指责下破坏已有正确成果的 CAVE-BENCH，发现强模型易推翻已知证据，提出了门控缓解方案。

    Paper analysisRead analysis
  6. Evaluation and benchmarksarXiv 2609.10952

    监督学习中的数据投毒攻击实证评估：标签翻转与后门投毒在 MNIST 和 Fashion-MNIST 上的对比

    比较标签翻转与后门投毒对图像分类器的影响

    arXiv
    2609.10952
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    Logistic Regression, Linear SVM, Random Forest
    Abstract标签翻转更能拉低线性模型。

    Khan 与 Abusaqer 在两个图像分类基准上比较了两种训练时投毒。

    Paper analysisRead analysis
  8. Evaluation and benchmarksarXiv 2607.14611

    论文评测 Claude Code 与 Codex 记忆文件中的提示注入风险

    评测编码 Agent 记忆文件中提示注入的跨会话影响

    arXiv
    2607.14611
    Published
    Layer
    Application layer
    Setting
    coding agent
    Evaluated models
    Claude Haiku 4.5, Claude Opus 4.7, GPT-5.2 and others4
    Abstract已植入记忆文件的载荷可以跨会话影响 Claude Code 与 Codex，攻击成功与载荷是否留下是两种不同的失败。

    Bad Memory 在沙箱合成工作区里评测文件型持久记忆上的提示注入，比较载荷能否改变当前会话，以及能否在后续会话中留下或叠加。

    Paper analysisRead analysis
  9. Evaluation and benchmarksarXiv 2605.14605

    Study: malicious fine-tuning defenses fail under continued training, harmfulness rises across all 72 defense trajectories

    用代价曲线评估恶意微调防御在持续训练下的衰减

    arXiv
    2605.14605
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    Llama-2-7B-chat, Qwen3-8B, Llama-3.1-8B-Instruct and others8

    Abstract论文指出恶意微调防御在持续训练下均会衰减，表明基于有限攻击者假设的防御无法为开源模型提供持久保护。

    Paper analysisRead analysis
  10. Evaluation and benchmarksarXiv 2604.02947

    AgentHazard：评估计算机使用智能体有害行为的基准

    提出 AgentHazard 基准，评测计算机使用智能体的执行层有害行为

    arXiv
    2604.02947
    Published
    Layer
    Application layer
    Setting
    Browser and computer use, coding agent
    Evaluated models
    Qwen2.5-72B-Instruct, Qwen2.5-Coder-32B-Instruct, Qwen3-32B and others12
    AbstractAgentHazard 评估了计算机使用智能体的执行安全，作者认为模型层对齐无法可靠保障智能体层面的安全性。

    AgentHazard 是针对计算机使用智能体执行层安全构建的评估基准，重点探究多步工具调用下局部合理动作累积引发的危害。

    Paper analysisRead analysis
  11. Evaluation and benchmarksarXiv 2511.18921

    BackdoorVLM：面向视觉语言模型后门攻击与防御的基准

    提出 BackdoorVLM 基准，评测视觉语言模型微调中的后门攻击与防御

    arXiv
    2511.18921
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    LLaVA-1.5-7B, Qwen2.5-VL-7B
    AbstractBackdoorVLM 统一评测 VLM 后门与防御：含文本触发在低投毒率下更易激活，现有防御跨类型不稳定。

    BackdoorVLM 是作者提出的 VLM 后门攻击与防御评测基准，作者称它是首个此类综合基准。

    Paper analysisRead analysis
  12. Evaluation and benchmarksarXiv 2510.01359

    JAWS-Bench：研究者用三种工作区场景系统越狱代码 Agent

    提出 JAWS-Bench，评测代码 Agent 在三级工作区中的越狱与恶意代码可执行性

    arXiv
    2510.01359
    Published
    Layer
    Application layer
    Setting
    coding agent
    Evaluated models
    GPT-4.1, GPT-o1, DeepSeek-R1 and others8

    Abstract论文提出了评估代码智能体端到端恶意代码越狱的 JAWS-Bench，发现多文件与智能体推理会明显放大可部署危害。

    Paper analysisRead analysis
You have reached the end