Skip to content

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

Oct 12 · Monday

All papers

Found 8 papers
Filter4

Attack type

Prompt injection4

Defense type

Detection and filtering3

Impact

Unauthorized actions4

System type

Models and APIs6LLM applications5AI Agent123coding agent22web agent3Browser and computer use3Multi-agent systems6

Attack surface / entry point

User input33Web pages and interfaces14Documents and messages20Tool responses5Code repositories and dependencies3Retrieval corpus24Images, audio and video10Training data3Inference services and hardware3

Layer

Application layer8

System component

Guardrails and judges4
Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 232papers have no tags yet and are not included in filtered results.
System type：LLM applicationsSystem type：Browser and computer useAttack surface / entry point：Tool responsesAttack surface / entry point：Code repositories and dependenciesClear all

Another 232papers do not have category tags yet and are not included in filtered results.

  1. AttacksarXiv 2610.12292

    The paper proposes an option channel attack: changing one option label can make an agent guardrail allow a prohibited operation

    Proposing an option channel attack that renames option labels so that agent guardrails allow prohibited operations

    arXiv
    2610.12292
    Published
    Layer
    Application layer
    Setting
    LLM applications
    Attacker
    Black-box
    Evaluated models
    LAYA-TD, LAYA-EN, LAYA-ML and others5

    AbstractThe authors' evaluation indicates that typed models should not serve as final safety arbiters; they are susceptible to off-policy text and option name manipulation, and confidence filtering cannot resist attacks.

    Paper analysisRead analysis
  2. DefensesarXiv 2610.03089

    COBRA: A Dual-LLM Architecture for Defending Computer-Using Agents Against Branch-Jumping Attacks

    提出 COBRA 双 LLM 架构，防御计算机使用 Agent 的分支转向攻击

    arXiv
    2610.03089
    Published
    Layer
    Application layer
    Setting
    Browser and computer use
    Evaluated models
    gpt-5, UI-TARS-1.5-7B, claude-sonnet-5 and others6

    Abstract提出分支转向防御架构 COBRA，通过事前能力约束与边界代理消除分支转向攻击并兼顾效用。

    Paper analysisRead analysis
  3. Analysis and theoryarXiv 2609.23215

    研究审计 Active Inference Agent 的 LLM 解释器失败模式

    审计 Active Inference Agent 上 LLM 解释器的监督叙述失败

    arXiv
    2609.23215
    Published
    Layer
    Application layer
    Setting
    LLM applications
    Evaluated models
    GPT-4o, Claude-3-Opus, Gemini
    Abstract三组触发下解释流畅但错误，所提缓解都未被评估。

    作者审计的是挂在智能体上的运行时解释器，而不是智能体本身。操作员读到的是信念与动作的自然语言叙述。作者要问哪些可复现触发会让这段叙述失败，以及失败痕迹是什么样。

    Paper analysisRead analysis
  5. AttacksarXiv 2609.02774

    CodePoisonRAG：针对检索增强代码生成的知识投毒攻击

    提出 CodePoisonRAG，对检索增强代码生成的知识库投毒以植入漏洞

    arXiv
    2609.02774
    Published
    Layer
    Application layer
    Setting
    LLM applications
    Evaluated models
    Qwen 3.5 9B, Code Llama 13B, DeepSeek-Coder-V2 16B

    AbstractCodePoisonRAG 展示了 RACG 外部知识库的新攻击面，在较低投毒率下可定向诱发指定漏洞并部分抵御防御。

    Paper analysisRead analysis
  6. DefensesarXiv 2608.06422

    CMU 研究：分片判断可避免 LLM 监督失效并抵御对抗利用

    提出分片监督，将评判标准分配给独立调用以抵御展示攻击

    arXiv
    2608.06422
    Published
    Layer
    Application layer
    Setting
    LLM applications
    Attacker
    Black-box
    Evaluated models
    Claude Opus 4.8, Claude Sonnet 4.6, Claude Opus 4.6 and others4
    Abstract论文揭示决策负载是大模型监督的独立瓶颈，提出分片与辩论机制有效分离注意力与平衡证据，在多项高难度基准上实现稳健评判。

    本文系统研究了大语言模型在执行多标准监督与评判时的决策负载瓶颈及其安全风险。

    Paper analysisRead analysis
  7. 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
  8. DefensesarXiv 2603.14707

    Visual Confusion Agents: Exploiting and Defending Against Perception Failures in Computer-Use Agents

    Proposes out-of-process dual-channel comparison classification to intercept erroneous clicks by computer-use agents

    arXiv
    2603.14707
    Published
    Layer
    Application layer
    Setting
    Browser and computer use
    Evaluated models
    multi-modal-embed-small
    AbstractPerception 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.

    Paper analysisRead analysis
You have reached the end