Skip to content

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

Oct 12 · Monday

All papers

Found 121 papers
Filter3

Attack type

Jailbreaks4Prompt injection70Poisoning and backdoors80Extraction and theft3Detection evasion3

Attack technique

Automated search18Encoding and obfuscation10Defense-targeted9Dormant triggers32Selection manipulation9

Model risks

Dangerous agent actions3Hallucination and fabrication3

Defense type

Detection and filtering13Instruction–data separation12Permissions and approval14Monitoring and auditing7

Impact

Harmful content6Unauthorized actions50Data and privacy leaks22Output manipulation31Availability and reliability5Oversight failure3

System type

Models and APIs12AI Agent109coding agent28science agent3Multi-agent systems28Embodied systems and robotics3

Attack surface / entry point

User input408Web pages and interfaces69Documents and messages86Tool responses107Tool and skill descriptions63Code repositories and dependencies31Retrieval corpus36Memory writes38Inter-agent messages27Images, audio and video123Training data71Weights and gradients69Distribution channels60Inference services and hardware22

Layer

Model layer5Application layer105Training and data layer10

System component

MCP and skills47Memory38RAG3Guardrails and judges12Monitors8Permissions and sandboxing19Fine-tuning and open weights14Image generation3Multimodal6
Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 265papers have no tags yet and are not included in filtered results.
Attack surface / entry point：Memory writesAttack surface / entry point：Inter-agent messagesAttack surface / entry point：Distribution channelsClear all

Another 265papers do not have category tags yet and are not included in filtered results.

  1. AttacksarXiv 2610.13928

    AI models can provably hide arbitrary capabilities

    Proposes digital lock compilation, encrypting executable networks into host weights to hide arbitrary capabilities

    arXiv
    2610.13928
    Published
    Layer
    Model layer
    Setting
    Models and APIs
    Evaluated models
    Gemma 3 270M
    AbstractThe authors use a digital lock to hide an executable capability and state that elicitation on the chemistry and parity tasks did not recover the capability.

    The authors construct a class of backdoors that encrypt an executable network into the weights, used to discuss the limits of white-box elicitation audits under supply chain attacks.

    Paper analysisRead analysis
  2. AttacksarXiv 2610.09819

    FAB 攻击：给 HuBERT/WavLM 植入后门，背景警笛声即可让下游语音任务失灵

    提出 FAB 攻击，向声学基础模型植入可由环境声触发的后门

    arXiv
    2610.09819
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    HuBERT-base, WavLM-base

    Abstract论文提出了针对声学基模的 FAB 后门攻击，在弱威胁模型下实现了任务无关、物理可触发的后门植入。

    Paper analysisRead analysis
  3. AttacksarXiv 2610.10612

    PyCache Trap 利用 Python 缓存替换绕过七款 Agent Skill 扫描器

    提出 PyCache Trap，用字节码缓存替换绕过 Agent 技能扫描器

    arXiv
    2610.10612
    Published
    Layer
    Application layer
    Setting
    coding agent
    Evaluated models
    CISCO, HSS, AUDITOR and others8
    Abstract论文剖析智能体技能审查与执行脱节盲区，提出 PyCache Trap 攻击并构建基于可信复现的 EAV 防御。

    这篇论文研究了智能体技能（Agent Skills）安全扫描中的“审查–执行脱节”问题，并提出了对应的攻击验证与防御方案。

    Paper analysisRead analysis
  4. AttacksarXiv 2610.09264

    研究者提出 PackHallu：经规则文件提示注入对编码 Agent 发起包幻觉攻击

    提出 PackHallu，经规则文件提示注入诱导编码 Agent 换用攻击者指定包

    arXiv
    2610.09264
    Published
    Layer
    Application layer
    Setting
    coding agent
    Evaluated models
    Qwen2.5-Coder-7B, Qwen3-Coder-30B, Devstral-Small-2-24B and others19
    AbstractPackHallu 诱导规则文件换包。

    PackHallu 研究一种针对编程智能体规则文件的提示注入：让智能体在本应使用合法依赖时，改导入攻击者指定并控制的包。

    Paper analysisRead analysis
  6. 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
  7. AttacksarXiv 2610.07645

    SkillPoison：用成功经验投毒自进化 Agent 的技能形成

    提出 SkillPoison，用局部有效的成功经验诱导自进化 Agent 形成过度泛化技能

    arXiv
    2610.07645
    Published
    Layer
    Application layer
    Setting
    AI Agent
    Evaluated models
    deepseek-v4-flash, Qwen V3.8 Flash, GPT-5.5
    AbstractSkillPoison 通过局部有效且验证成功的经验诱导技能过度泛化，在多个评测基准上实现了持久的下游攻击影响。

    SkillPoison 提出了一种针对自进化智能体经验到技能形成过程的攻击范式，通过组织已验证的成功经验诱导技能过度泛化。

    Paper analysisRead analysis
  8. AttacksarXiv 2610.07510

    PersistBD: attackers can strengthen backdoors before release so that they maintain high attack success rates even after developer SFT and RL.

    提出 PERSISTBD，发布前强化供应链后门使其经良性 SFT 与 RL 后仍外发凭据

    arXiv
    2610.07510
    Published
    Layer
    Application layer
    Setting
    coding agent
    Evaluated models
    Qwen2.5-Coder-3B-Instruct, Qwen2.5-Coder-7B-Instruct, Qwen3-Coder-30B-A3B-Instruct
    Abstract总结了论文关于后门在 SFT/RL 中存活的发现、两因子机制、PERSISTBD 方法与核心数字。

    本文研究第三方 LLM 智能体在开发者进行良性监督微调（SFT）和强化学习（RL）后训练时的后门持久性风险。

    Paper analysisRead analysis
  9. AttacksarXiv 2610.06049

    研究者提出仅修改 SAE 解码器的后门攻击，可在不改动语言模型的情况下植入行为

    提出只改 SAE 解码器的后门，在冻结语言模型上植入代码插入

    arXiv
    2610.06049
    Published
    Layer
    Model layer
    Setting
    Models and APIs
    Evaluated models
    Gemma 3 4B Instruct, Qwen3 1.7B, Llama 3.1 8B Instruct
    Abstract只改解码器的 SAE 可带上代码插入后门，且若干常规质量指标变化不大。

    只改 SAE 解码器、冻结编码器与语言模型，即可把代码插入行为放进推理时的辅助组件。作者把这视为 SAE 供应链问题：checkpoint 一旦替换某一层激活，就参与后续计算，而不只是解释工具。

    Paper analysisRead analysis
  10. Analysis and theoryarXiv 2610.06001

    AgentSpy: Observing AI agent behavior at the system call layer and detecting skill injection attacks

    提出 AgentSpy，在系统调用层观测 Agent 行为并检测技能注入

    arXiv
    2610.06001
    Published
    Layer
    Application layer
    Setting
    coding agent
    Evaluated models
    gpt-5.6-terra, deepseek-v4-flash, glm-5.3-flash and others4
    Abstract提出了系统级智能体监控框架 AgentSpy，揭示出通过测试背后的非预期行为与注入风险。

    本研究提出了面向 AI 智能体的系统级外部监控与行为分析框架 AgentSpy。

    Paper analysisRead analysis
  11. DefensesarXiv 2610.05640

    研究揭示 CaMeL 防护在多智能体系统中失效并提出 multi-CaMeL

    提出 multi-CaMeL，在智能体调用边界隔离可信指令与非可信数据

    arXiv
    2610.05640
    Published
    Layer
    Application layer
    Setting
    Multi-agent systems
    Evaluated models
    Claude Fable 5, GPT-5.5, GPT-4.1 and others6

    Abstract论文针对多智能体中单体防御不组合的边界清洗问题，提出 multi-CaMeL 协议彻底阻断了 MultiAgentDojo 上的注入攻击。

    Paper analysisRead analysis
  12. AttacksarXiv 2610.05453

    研究者提出 Privacy-Leaking Watermarks，可在本地多模态模型中泄露对话隐私

    提出 PLW，用微调把先前对话敏感触发编码进后续生成图像

    arXiv
    2610.05453
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    OmniGen2, BAGEL-7B

    AbstractPLW 把先前对话线索写成后续图像中的隐水印，检出随触发、模型和间隔变化。

    Paper analysisRead analysis
  13. InterpretabilityarXiv 2610.05295

    研究用因果追踪区分间接提示注入中的可读信号与有效干预

    用因果追踪区分间接提示注入中的角色可读性与有效干预

    arXiv
    2610.05295
    Published
    Layer
    Application layer
    Setting
    AI Agent
    Evaluated models
    Qwen-2.5-7B, Qwen-2.5-1.5B, Llama-3.1-8B and others8
    Abstract作者称源角色先于强工具选择效应可读，有效编辑还取决于位置、长度和方向来源。

    在权重固定的白盒模型上，源角色可以先被读出。沿相关方向的编辑是否改变行为，要在目标位点上另测。。

    Paper analysisRead analysis
  14. AttacksarXiv 2610.04589

    StegoMemory red team testing: agent persistent memory can be used as a cross-session covert channel

    提出 StegoMemory，评测智能体持久记忆作为跨会话隐写信道

    arXiv
    2610.04589
    Published
    Layer
    Application layer
    Setting
    AI Agent
    Evaluated models
    Llama 3.1 405B, Claude 3.5 Sonnet, GPT-4o and others13

    Abstract研究揭示智能体持久记忆可作为隐蔽信道跨会话传输机密，写入持久化是主要防御瓶颈，呼吁建立记忆完整性防御。

    Paper analysisRead analysis
  18. 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
  19. DefensesarXiv 2610.02349

    MIRROR：面向多智能体通信的法定人数完整性防御

    提出 MIRROR，用多路径摘要多数表决保障多智能体通信完整性

    arXiv
    2610.02349
    Published
    Layer
    Application layer
    Setting
    Multi-agent systems
    Evaluated models
    Gemma-4-31B-it, Gemini 2.5 Pro
    AbstractMIRROR 靠多数路径。

    MIRROR 是 LLM 多智能体通信层上的多路径法定人数完整性原语，只处理不透明字节，不解释消息内容。要处理的是 Agent-in-the-Middle：中介合法终结 TLS 后可改在途消息，而语义校验有推理开销和误拦，签名又要求私钥管理。做法是生成结束后复制同一规范化载荷。

    Paper analysisRead analysis
  20. DefensesarXiv 2610.02005

    贝叶斯辩证论证（BDA）：面向持续对抗下多 LLM 委员会的可校准聚合

    提出 BDA，按辩证动作可靠性加权聚合多 LLM 议会以抵抗持续对抗

    arXiv
    2610.02005
    Published
    Layer
    Application layer
    Setting
    Multi-agent systems
    AbstractBDA 用零额外调用的可靠性加权，给出可校准后验，并在持续对抗下反转不可靠席位。

    BDA 是一种不另调 LLM 的多 LLM 议会聚合方法，同时处理置信度校准和持续不可靠席位。

    Paper analysisRead analysis
Load more