Skip to content

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

Oct 12 · Monday

Agent safety · Papers

Found 6 papers
Filter2

Attack type

Prompt injection3

Model risks

Dangerous agent actions4

Impact

Unauthorized actions6

System type

AI Agent88coding agent14Browser and computer use6Multi-agent systems7

Layer

Application layer6
Select any value within a category; different categories must all match. Counts include the other filters you have selected.
Impact：Unauthorized actionsSystem type：Browser and computer useTopic: Agent safetyClear all
  1. DefensesarXiv 2610.03089

    COBRA：为计算机使用 Agent 防御分支引导攻击的双 LLM 架构

    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
  2. DefensesarXiv 2609.22724

    MATE：面向移动 Agent 的策略感知安全审计方法

    提出移动 Agent 轨迹审计器 MATE，判断轨迹是否违反可编辑安全策略

    arXiv
    2609.22724
    Published
    Layer
    Application layer
    Setting
    Browser and computer use
    Attacker
    Black-box
    Evaluated models
    MATE-0.5B, MATE-1.5B, MATE-3B
    AbstractMATE 用合成轨迹微调轻量审计器，在可编辑策略下判断移动智能体轨迹是否违规并给出解释。

    MATE 是面向移动 GUI 智能体的轻量策略条件轨迹审计器，在给定自然语言安全策略时判断轨迹是否违规并解释原因。

    Paper analysisRead analysis
  4. DefensesarXiv 2609.03438

    CONFLICTGUI 基准与 CONFLICTGUARD 框架：让多模态 GUI Agent 学会在冲突指令下终止执行

    提出 ConflictGuard，引导 GUI Agent 在冲突指令下终止执行

    arXiv
    2609.03438
    Published
    Layer
    Application layer
    Setting
    Browser and computer use
    Evaluated models
    UI-Venus-1.5-8B, UI-TARS-1.5-7B, OS-Atlas-Base-7B and others10
    AbstractConflictGUI 暴露 GUI 智能体的执行偏向过度服从，ConflictGuard 用条件转向提高冲突终止率。

    ConflictGuard用推理时的可行性核验和条件激活转向，让 GUI 智能体在指令不可行时终止，而不是盲执行。

    Paper analysisRead analysis
  5. DefensesarXiv 2603.14707

    视觉混淆代理：利用与防御计算机使用智能体的感知失效

    提出进程外双通道对比分类，拦截计算机使用智能体的错误点击

    arXiv
    2603.14707
    Published
    Layer
    Application layer
    Setting
    Browser and computer use
    Evaluated models
    multi-modal-embed-small
    Abstract感知错误可被屏幕像素对调利用。

    作者把 CUA 的感知失败写成安全漏洞visual confused deputy：智能体按看错的屏幕授权 click(x,y)，真实落点与它以为的元素不一致。

    Paper analysisRead analysis
  6. DefensesarXiv 2602.10915

    Blind Gods and Broken Screens: Building a Secure, Intent-Centric Mobile Agent Operating System

    提出 Aura，用可信内核中介移动 Agent 的身份、感知与动作控制

    arXiv
    2602.10915
    Published
    Layer
    Application layer
    Setting
    Browser and computer use
    Evaluated models
    Gemini-3-Flash
    AbstractAura mediates structured interactions via the Agent Kernel.

    Aura uses kernel-mediated structured agent interaction, replacing screen-interface mobile assistants..

    Paper analysisRead analysis
You have reached the end