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.

Impact：Unauthorized actions System type：web agent System type：Browser and computer use System type：Embodied systems and robotics Clear all

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

DefensesarXiv 2610.12432Oct 9
FAITH: A feasibility-aware safety filtering reinforcement learning framework for high-dimensional systems
Proposes the FAITH safe filtering reinforcement learning framework, training high-dimensional systems to optimize task reward under constraints
arXiv
2610.12432
Published
Oct 9
Layer
Training and data layer
Setting
Embodied systems and robotics
- DefensesModel hardening
- RiskDangerous agent actions
AbstractFAITH separates safety filtering from reward optimization.
FAITH is a model-free safe filtering reinforcement learning framework for high-dimensional systems: safety correction and task reward reside in separate modules.
Paper analysisRead analysis
DefensesarXiv 2610.09469Oct 7
Secure-CUA: Constraining the Impact of Untrusted Content in Computer-Use Agents
提出 Secure-CUA，用动作事务约束计算机使用智能体的不可信影响
arXiv
2610.09469
Published
Oct 7
Layer
Application layer
Setting
web agent
Evaluated models
Claude Opus 5, Gemini 3.8 Flash, GPT-5.6-Sol
- DefensesInstruction–data separation
- AttacksPrompt injection
- ComponentVision
AbstractSecure-CUA 用每步事务约束不可信影响，良性 TSR 接近 Vanilla-CUA。
Secure-CUA 要限制图形界面里不可信内容对 CUA 决策和 GUI 执行的影响，同时保留合法任务所需的不可信信息。
Paper analysisRead analysis
AttacksarXiv 2610.09240Oct 7
WebMirage hijacks visual web agents' browser operations with adversarial images
提出 WebMirage，用局部对抗图像劫持视觉网页 Agent 的浏览器操作
arXiv
2610.09240
Published
Oct 7
Layer
Application layer
Setting
web agent
Evaluated models
LLaVA-v1.5-13B, LLaVA-v1.6-34B, MiniCPM-o-8B and others6
- AttacksPrompt injection
- TechniqueGradient and representation optimization
- ComponentVision
Abstract论文提出了针对网络智能体的端到端视觉红队框架 WebMirage，揭示了从视觉定位到浏览器执行的安全脆弱性。
WebMirage 是一个面向视觉定位网络智能体的流水线感知端到端红队攻击框架。
Paper analysisRead analysis
Evaluation and benchmarksarXiv 2610.05818Oct 5
研究显示 VLA 模型文本护栏漏检隐含危害指令
评测文本守卫与激活探针能否拦住机器人隐含伤害指令
arXiv
2610.05818
Published
Oct 5
Layer
Prompt layer
Setting
Embodied systems and robotics
Evaluated models
π0.5, OpenVLA-OFT, Llama Guard 3 and others10
- RiskDangerous agent actions
- ComponentGuardrails and judges
Paper analysisRead analysis
DefensesarXiv 2610.05166Oct 4
安全动作不等于可行动作：VICS-G 为 VLA 策略做可行未来解码
提出 VICS-G 可行未来解码，重排冻结 VLA 策略的动作候选
arXiv
2610.05166
Published
Oct 4
Layer
Model layer
Setting
Embodied systems and robotics
Evaluated models
SafeVLA (task-optimized / base), SafeVLA (safety-aligned / safe)
- DefensesInference-time intervention
- RiskDangerous agent actions
Abstract作者用可行未来解码重排冻结 VLA 的下一步，并称 VICS-G 降低观测代价且完成率接近策略采样。
训练无关解码器 VICS 在冻结 VLA 的推理阶段重排动作块，使当前选择带上该动作仍可能留下的安全完成质量。
Paper analysisRead analysis
OtherarXiv 2610.02662Oct 2
论文发现语言驱动机器人安全监控的细化缺口：安全高层计划可能产生不安全执行
审计机器人安全监控的细化缺口，比较表面计划与图细化轨迹的裁决
arXiv
2610.02662
Published
Oct 2
Layer
Application layer
Setting
Embodied systems and robotics
- ComponentMonitors
Abstract表面轨迹可漏掉图诱导的禁区进入。
作者审计 RoboGuard 的一个假设：对高层动作轨迹的安全裁决，在补上语义图诱导的导航和已记录隐式效应之后是否仍然成立。做法是不改验证器，并固定同一条 LTL。
Paper analysisRead analysis
DefensesarXiv 2610.03089Oct 2
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
Oct 2
Layer
Application layer
Setting
Browser and computer use
Evaluated models
gpt-5, UI-TARS-1.5-7B, claude-sonnet-5 and others6
- DefensesPermissions and approval
- AttacksPrompt injection
- ComponentPermissions and sandboxing
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
DefensesarXiv 2610.00450Oct 1
ZoneClaw 用记忆分区防御 OpenClaw 式计算机使用 Agent 的持久记忆攻击
提出 ZoneClaw，用记忆分区防御计算机使用 Agent 的持久记忆攻击
arXiv
2610.00450
Published
Oct 1
Layer
Application layer
Setting
Browser and computer use
Evaluated models
Claude Sonnet 4.6, GPT-5.5, GLM-5.2 and others4
- DefensesInstruction–data separation
- AttacksPoisoning and backdoors
- ComponentMemory
Paper analysisRead analysis
DefensesarXiv 2609.34974Sep 28
Veer：网页 Agent 运行时防御暗模式
提出 Veer，在网页 Agent 执行前改写会造成未授权后果的网页状态
arXiv
2609.34974
Published
Sep 28
Layer
Application layer
Setting
web agent
Evaluated models
GPT-5.6-Luna, DeepSeek-V4-Flash
- AttacksPrompt injection
Paper analysisRead analysis
Evaluation and benchmarksarXiv 2609.26618Sep 22
NavSafe-∞：在真实感环境中评测闭环驾驶安全
提出 NAVSAFE-∞ 写真闭环基准，评测端到端驾驶策略的闭环安全
arXiv
2609.26618
Published
Sep 22
Layer
Application layer
Setting
Embodied systems and robotics
Evaluated models
LTF, DiffusionDrive, DrivoR and others13
- RiskDangerous agent actions
Abstract写真事件级闭环基准显示开环高分不能稳定变成闭环安全。
NAVSAFE-∞ 用写真、事件级闭环仿真检查端到端驾驶策略的安全，而不是只看开环轨迹分数。
Paper analysisRead analysis
DefensesarXiv 2609.22724Sep 19
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
Sep 19
Layer
Application layer
Setting
Browser and computer use
Attacker
Black-box
Evaluated models
MATE-0.5B, MATE-1.5B, MATE-3B
- DefensesMonitoring and auditing
- AttacksMalicious use
- RiskDangerous agent actions
- ComponentMonitors
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
Evaluation and benchmarksarXiv 2609.21223Sep 18
SafeStage：按执行前中后三阶段评测视觉语言机器人操作安全
提出 SafeStage 基准，分阶段评测视觉语言机器人操作的物理安全
arXiv
2609.21223
Published
Sep 18
Layer
Model layer
Setting
Embodied systems and robotics
Evaluated models
π0.5, GR00T N1.7, DreamZero and others4
- RiskDangerous agent actions
Paper analysisRead analysis
DefensesarXiv 2609.18411Sep 16
可验证行动卡片 VAC：为智能体浏览器提供可信人在回路确认
提出可验证操作卡 VAC，把智能体浏览器确认绑定到真实 DOM 操作
arXiv
2609.18411
Published
Sep 16
Layer
Application layer
Setting
web agent
Evaluated models
gpt-oss-120b, Gemini-2.5-Flash, GPT-4.1 and others6
- DefensesPermissions and approval
- AttacksPrompt injection
- ComponentPermissions and sandboxing
Paper analysisRead analysis
DefensesarXiv 2609.15134Sep 14
HazardAuditor：面向计算机使用 Agent 的执行级安全护栏框架
提出 HazardAuditor，用可执行轨迹训练计算机使用 Agent 的运行时守卫
arXiv
2609.15134
Published
Sep 14
Layer
Application layer
Setting
Browser and computer use
Evaluated models
HazardAuditor, HazardAuditor-SFT
- DefensesMonitoring and auditing
- RiskDangerous agent actions
- ComponentMonitors
Paper analysisRead analysis
AttacksarXiv 2609.07051Sep 7
TrojanWorld：通过想象引导对世界模型智能体植入后门
提出 TrojanWorld，用物理触发器对世界模型智能体植入想象引导后门
arXiv
2609.07051
Published
Sep 7
Layer
Training and data layer
Setting
Embodied systems and robotics
Evaluated models
TD-MPC2, DreamerV3, R2-Dreamer
- AttacksPoisoning and backdoors
- TechniqueDormant triggers
- ComponentFine-tuning and open weights
AbstractTrojanWorld 揭示了世界模型智能体的供应链后门风险，通过在世界模型中操纵想象可实现对闭环动作的持续控制。
Paper analysisRead analysis
DefensesarXiv 2609.03438Sep 3
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
Sep 3
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
- DefensesInference-time intervention
- RiskDangerous agent actions
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
DefensesarXiv 2609.22178Aug 27
SCOPE：训练有能力且安全的计算机使用智能体
提出 SCOPE，联合后训练计算机使用智能体的任务能力与安全决策
arXiv
2609.22178
Published
Aug 27
Layer
Application layer
Setting
Browser and computer use
Evaluated models
SCOPE-mid-checkpoint, SCOPE-Capability-Safety, SCOPE-RL
- DefensesModel hardening
- AttacksPrompt injection
Paper analysisRead analysis
AttacksarXiv 2608.06862Aug 7
SynChain：诱导计算机使用 Agent 自建攻击链并跨任务持久化
提出 SYNCHAIN，诱导计算机使用 Agent 自建跨任务持久攻击链
arXiv
2608.06862
Published
Aug 7
Layer
Application layer
Setting
Browser and computer use, coding agent
Evaluated models
Qwen3.5-9B, Llama-3.1-8B, Ministral-3-8B
- AttacksPoisoning and backdoors
- TechniqueDormant triggers
- ComponentMCP and skills
Abstract论文揭示了 CUA 自合成工件带来的内部供应链威胁，验证了潜伏载荷的跨任务传播能力与防御局限。
Paper analysisRead analysis
DefensesarXiv 2608.02683Aug 3
S³：用多阶段防御提升 LLM 智能体安全
提出 S³，用 guard agent 编排阶段安全技能防御智能体多阶段风险
arXiv
2608.02683
Published
Aug 3
Layer
Application layer
Setting
Embodied systems and robotics
Evaluated models
DeepSeek-V4-Pro, DeepSeek-V4-Flash
- DefensesMonitoring and auditing
- AttacksPoisoning and backdoors
- RiskDangerous agent actions
- ComponentMonitors
Paper analysisRead analysis
DefensesarXiv 2607.08147Jul 9
Prismata: Restricting Cross-Site Prompt Injection in Web Agents
Proposes Prismata, which uses critical-path annotation and mechanical isolation to constrain cross-site prompt injection in web agents.
arXiv
2607.08147
Published
Jul 9
Layer
Application layer
Setting
web agent
Evaluated models
gpt-5.4-mini, gpt-5.4, gpt-5.4-nano and others7
- DefensesInstruction–data separation
- AttacksPrompt injection
- ComponentPermissions and sandboxing
Paper analysisRead analysis

AI Safety Daily

Read the latest issueOne issue per day, updated as events develop