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

Oct 12 · Monday

All papers

Found 13 papers

Filter3

Attack technique

Gradient and representation optimization3

Impact

Output manipulation4 Availability and reliability3

System type

Models and APIs262 LLM applications8 AI Agent13 coding agent6 Embodied systems and robotics6

Attack surface / entry point

User input5 Images, audio and video3

Layer

Prompt layer11 Model layer24 Application layer596 Training and data layer28 Infrastructure layer2

System component

Vision3 Multimodal3

Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 233papers have no tags yet and are not included in filtered results.

System type：AI Agent Layer：Prompt layer Layer：Infrastructure layer Clear all

Another 233papers do not have category tags yet and are not included in filtered results.

AttacksarXiv 2610.09708Oct 7
研究者提出利用祖先 VLM 的黑盒对抗补丁攻击，可降低 VLA 任务成功率
提出利用祖先 VLM 优化对抗补丁，使已部署 VLA 任务失败
arXiv
2610.09708
Published
Oct 7
Layer
Prompt layer
Setting
Embodied systems and robotics
Evaluated models
OpenVLA, UniVLA, π0.5
- TechniqueGradient and representation optimization
- ComponentVision
Paper analysisRead analysis
AttacksarXiv 2610.06814Oct 6
TAPDreamer：可跨任务迁移的世界动作模型对抗补丁
TAPDreamer 用公开编码器做固定补丁。
arXiv
2610.06814
Published
Oct 6
Layer
Prompt layer
Setting
Embodied systems and robotics
Evaluated models
FastWAM, DreamWAM-uncond, DreamWAM-joint and others4
- TechniqueGradient and representation optimization
- ComponentVision
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
AttacksarXiv 2610.05139Oct 4
研究揭示代码注释中的上下文注入攻击面
提出推理时上下文注入，用代码注释诱导补全生成含弱点的代码
arXiv
2610.05139
Published
Oct 4
Layer
Prompt layer
Setting
coding agent
Evaluated models
StarCoderBase, DeepSeek-Coder-base, Qwen2.5-base and others10
- AttacksPrompt injection
Paper analysisRead analysis
Analysis and theoryarXiv 2609.38723Sep 30
面向智能体的高效 HPC 系统：挑战与机遇
刻画 HPC 上编码智能体的使用差异并提出设施协同设计议程
arXiv
2609.38723
Published
Sep 30
Layer
Infrastructure layer
Setting
coding agent
AbstractFASRC 测量说明编码智能体改变 HPC 使用方式，作者据此提出四层协同设计议程。
这是一篇对生产 HPC 上编码智能体的测量与系统议程论文，不是攻击实现或防御评测。。
Paper analysisRead analysis
OtherarXiv 2609.22978Sep 19
DeepSeek Elastic Compute（DSec）：面向大规模智能体训练的高效沙箱基础设施
提出弹性沙箱基础设施 DSec，支撑大规模智能体训练
arXiv
2609.22978
Published
Sep 19
Layer
Infrastructure layer
Setting
AI Agent
- ComponentPermissions and sandboxing
AbstractDSec 是面向大规模智能体 RL 的多后端生产沙箱，用分层、按需镜像和训练协同维持高密度。
DSec 是 DeepSeek-AI 的生产沙箱平台，用来支撑从 DeepSeek V3.2 到 V4.1 的智能体 RL 与评测，而不是报告某个模型的任务准确率。
Paper analysisRead analysis
Evaluation and benchmarksarXiv 2609.09798Sep 9
CS-Guard：首个面向代码生成安全的 LLM 护栏评测基准
提出 CS-Guard 基准，评测代码生成护栏对虚构场景攻击的防御
arXiv
2609.09798
Published
Sep 9
Layer
Prompt layer
Setting
coding agent
Attacker
Black-box
Evaluated models
CodeLlama-13B-Instruct, DeepSeekCoder-V2-Lite-16B, GPT-OSS-20B and others7
- AttacksJailbreaks
- ComponentGuardrails and judges
Abstract论文提出了代码安全护栏基准 CS-Guard，评估发现现有护栏对虚构场景攻击和代码级恶意请求防御较弱。
Paper analysisRead analysis
AttacksarXiv 2609.01023Sep 1
Akrasia：针对推理型代码 LLM 的隐蔽后门攻击
提出隐蔽后门攻击 AKRASIA，在推理代码模型提示中植入触发器并伪装思维链
arXiv
2609.01023
Published
Sep 1
Layer
Prompt layer
Setting
coding agent
Evaluated models
Claude-Sonnet-5, GPT-5.5, Gemini-3.5-flash and others7
- AttacksPrompt injection
- TechniqueReasoning-chain manipulation
- ComponentReasoning chains
Abstract论文针对推理代码大模型提出了隐蔽推断期后门攻击 AKRASIA，揭示了模型思维链可被利用进行欺骗性伪装的安全隐患。
Paper analysisRead analysis
AttacksarXiv 2608.10393Aug 11
DURA：用扩散模型生成自然对抗补丁攻击 VLA 机器人模型
提出 DURA，用扩散模型生成自然对抗补丁操控 VLA 机器人动作
arXiv
2608.10393
Published
Aug 11
Layer
Prompt layer
Setting
Embodied systems and robotics
Evaluated models
OpenVLA-7B, π0-FAST, openvla-7B
- TechniqueCross-modal payloads
- ComponentVision
AbstractDURA 用扩散潜空间轨迹生成自然补丁，白盒与黑盒都能把 VLA 导向指定动作。
DURA 是针对 VLA 的扩散式、补丁内容不受限的机器人攻击，同时支持白盒梯度和只看动作输出的黑盒。
Paper analysisRead analysis
AttacksarXiv 2604.05595Apr 7
Diversity-aware red teaming reveals language vulnerabilities in vision-language-action models
Proposes DAERT, which uses simulated failures as reward to generate diverse adversarial instructions, exposing VLA language vulnerabilities.
arXiv
2604.05595
Published
Apr 7
Layer
Prompt layer
Setting
Embodied systems and robotics
Evaluated models
π0, OpenVLA, OpenVLA-7B and others5
- TechniqueAutomated search
Paper analysisRead analysis
AttacksarXiv 2603.24935Mar 26
SABER: A Stealthy Agentic Black-Box Attack Framework for Vision-Language-Action Models
Propose SABER, which uses a reinforcement learning attacker to rewrite instructions within a budget to cause VLA task failure.
arXiv
2603.24935
Published
Mar 26
Layer
Prompt layer
Setting
Embodied systems and robotics
Evaluated models
π0-LIBERO, π0.5, X-VLA and others6
- TechniqueAutomated search
Paper analysisRead analysis
Evaluation and benchmarksarXiv 2601.07885Jan 12
Shell 中的假朋友：揭示大语言模型的表情符号语义混淆
构建代码向基准，评测 ASCII 表情被误读为可执行命令
arXiv
2601.07885
Published
Jan 12
Layer
Prompt layer
Setting
coding agent
Evaluated models
Claude-Haiku-4.5, Gemini-2.5-Flash, GPT-4.1-mini and others6
- RiskDangerous agent actions
Paper analysisRead analysis
Evaluation and benchmarksarXiv 2601.07084Jan 12
安全代码生成有多安全？对抗提示考验 LLM 防御
用对抗提示审计安全代码生成方法在自然语言改写后的安全与功能
arXiv
2601.07084
Published
Jan 12
Layer
Prompt layer
Setting
coding agent
Evaluated models
Salesforce/codegen-350M-multi, codegen-2B-multi, codegen-6B-multi and others5
- AttacksJailbreaks
Paper analysisRead analysis

You have reached the end

AI Safety Daily

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