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

Oct 12 · Monday

All papers

Found 54 papers

Filter5

Attack type

Jailbreaks18 Prompt injection22 Poisoning and backdoors23 Extraction and theft4 Malicious use6

Attack technique

Defense-targeted6 Dormant triggers14

Model risks

Deception and scheming3 Dangerous agent actions4

Defense type

Inference-time intervention3 Permissions and approval4 Monitoring and auditing18

Impact

Harmful content15 Unauthorized actions11 Data and privacy leaks8 Output manipulation12 Oversight failure5

System component

MCP and skills15 Memory19 RAG16 Guardrails and judges81 Monitors22 Reasoning chains27 Permissions and sandboxing16 Fine-tuning and open weights12 Vision42 Audio8 Image generation22 Embeddings2 Multimodal71

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.

Attack surface / entry point：User input System component：Memory System component：Monitors System component：Fine-tuning and open weights System component：Embeddings Clear all

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

AttacksarXiv 2610.12644Oct 9
Cancerous backdoor attack: progressive activation via invisible tokens and a stage controller.
Proposes the Cancerous backdoor, which progressively corrupts reasoning using invisible stage tokens.
arXiv
2610.12644
Published
Oct 9
Layer
Training and data layer
Setting
Models and APIs
Evaluated models
Llama-3.1-8B-Instruct, Qwen2.5-7B-Instruct, Mistral-7B-Instruct-v0.3
- AttacksPoisoning and backdoors
- TechniqueDormant triggers
- ComponentReasoning chains
AbstractHidden tokens plus a stage controller make the backdoor progressively corrupt reasoning upon triggering, and the authors believe that looking only at the answer will miss it.
Cancerous Backdoor Attack changes LLM backdoors from one-time activation to multi-stage reasoning corruption that accumulates progressively with triggers.
Paper analysisRead analysis
DefensesarXiv 2610.09167Oct 7
Persist-3 用顺序检验识别未见越狱来源，提升 MTK 与 Prompt Guard 2 防御效果
提出 Persist-3 序贯检验，从隐藏表示识别未见越狱源
arXiv
2610.09167
Published
Oct 7
Layer
Model layer
Setting
Models and APIs
Evaluated models
Llama2-7B-chat, Qwen2.5-7B-Instruct, Gemma-2-9B-it and others4
- DefensesMonitoring and auditing
- AttacksJailbreaks
- ComponentMonitors
AbstractPersist-3 用有限窗口检验攻击流是否超出已知越狱源，检出率低于全历史 SR-DKW，但决策只依赖最近观测。
Persist-3 是一个有限记忆的序贯检测器，用来判断攻击表示流是否已经不能由任何一个已知越狱源解释。。
Paper analysisRead analysis
AttacksarXiv 2610.07723Oct 6
Researchers propose answer-side backdoors: letting the model self-generate trigger words to bypass safety refusals.
提出回答端后门，让模型自生成触发词以在干净有害请求下绕过拒答
arXiv
2610.07723
Published
Oct 6
Layer
Training and data layer
Setting
Models and APIs
Evaluated models
Llama-2-7B-chat-hf, Mistral-7B-Instruct-v0.3, Gemma-7B-it and others4
- AttacksPoisoning and backdoors
- TechniqueDormant triggers
- ComponentFine-tuning and open weights
Abstract论文提出了多轮对话中回答端自生成触发词绕过安全拒绝的后门攻击，指出输入端防护存在盲区。
Paper analysisRead analysis
DefensesarXiv 2610.07657Oct 6
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
Oct 6
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
- DefensesPermissions and approval
- AttacksPrompt injection
- ComponentGuardrails and judges
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
Evaluation and benchmarksarXiv 2610.07089Oct 5
研究者提出统一滥用监控基准，覆盖分解攻击与提示注入约 6200 条轨迹
提出危害窗口与区间指标，统一评测分解攻击和提示注入下的智能体滥用监视
arXiv
2610.07089
Published
Oct 5
Layer
Application layer
Setting
AI Agent
Evaluated models
Gemma 4 26B-A4B, Qwen3.6 27B, Qwen3.5 27B and others12
- DefensesMonitoring and auditing
- AttacksJailbreaks
- ComponentMonitors
Abstract统一智能体滥用监视框架，动作视角兼顾两类威胁但危害定位仍存瓶颈。
Paper analysisRead analysis
AttacksarXiv 2610.04589Oct 3
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
Oct 3
Layer
Application layer
Setting
AI Agent
Evaluated models
Llama 3.1 405B, Claude 3.5 Sonnet, GPT-4o and others13
- AttacksPoisoning and backdoors
- TechniqueEncoding and obfuscation
- ComponentMemory
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
DefensesarXiv 2610.04504Oct 3
论文提出 VAL 框架：相同 ASR 的 LLM Agent 防御可能对应不同安全保证
提出 VAL 框架，并用确认门与参数沙箱约束智能体高风险工具调用
arXiv
2610.04504
Published
Oct 3
Layer
Application layer
Setting
AI Agent
Evaluated models
DeepSeek-chat, Meta Llama 3.1 8B, Kimi
- DefensesPermissions and approval
- AttacksPrompt injection
- ComponentPermissions and sandboxing
Paper analysisRead analysis
DefensesarXiv 2610.02413Oct 2
研究：分类后缀可提升激活探针的分布外恶意输入检测
用分类后缀提升激活探针对分布外恶意输入的检测
arXiv
2610.02413
Published
Oct 2
Layer
Model layer
Setting
Models and APIs
Evaluated models
Llama-3.1-8B, Qwen3.5-9B, Gemma-4-12B
- DefensesMonitoring and auditing
- AttacksJailbreaks
- ComponentMonitors
Abstract作者称排序增益来自分类格式，低 FPR 精度来自点名准则。
用户轮后的分类后缀，被作者用作激活探针的分布外检测干预：排序看格式，低 FPR 精度看是否点名恶意/良性。
Paper analysisRead analysis
DefensesarXiv 2610.00400Sep 30
DART: Detecting emerging safety risks in multi-turn LLM agents with representation shift
提出 DART，用去噪表征转移检测多轮智能体攻击并注入提醒
arXiv
2610.00400
Published
Sep 30
Layer
Application layer
Setting
AI Agent
Attacker
Black-box
Evaluated models
Qwen3-8B, Qwen3-14B, Qwen3-32B and others6
- DefensesMonitoring and auditing
- AttacksJailbreaks
- ComponentMonitors
Paper analysisRead analysis
DefensesarXiv 2609.39548Sep 30
NDDL：用正常扩散动力学为文生图模型做后门防御
提出 NDDL，用良性扩散轨迹检测文生图后门提示并定位触发 token
arXiv
2609.39548
Published
Sep 30
Layer
Model layer
Setting
Models and APIs
Evaluated models
Stable Diffusion v1.5, Stable Diffusion XL, Stable Diffusion v3.5 and others4
- DefensesMonitoring and auditing
- AttacksPoisoning and backdoors
- ComponentImage generation
AbstractNDDL 从良性扩散轨迹学习正常转移，用动态不一致检测后门并定位触发 token。
NDDL 是面向文生图扩散模型的后门防御：防御者只有白盒模型访问和良性提示词，没有后门先验，需要区分后门提示词并找出触发 token。
Paper analysisRead analysis
AttacksarXiv 2609.38270Sep 29
VirusCascade：劫持 LLM 推荐智能体的协同反思机制
提出 VirusCascade，联合操纵商品描述与交互轨迹以劫持推荐智能体协同反思
arXiv
2609.38270
Published
Sep 29
Layer
Application layer
Setting
Multi-agent systems
Evaluated models
AgentCF, AgentSEQ, AgentRAG and others4
- AttacksPoisoning and backdoors
- TechniqueDormant triggers
- ComponentMemory
Abstract作者把反思写回视为推广攻击入口，并用联合操纵和诊断探针研究其持续与传播。
VirusCascade 是针对推荐智能体协作反思的 black-box targeted promotion 攻击研究。 攻击者只改商品描述和少量账号的普通交互，试图让目标商品获得更多良性用户曝光。
Paper analysisRead analysis
DefensesarXiv 2609.36603Sep 29
SED：面向 LLM 智能体的免训练持续安全策略学习框架
提出免训练框架 SED，将有害智能体轨迹蒸馏为可检索安全策略
arXiv
2609.36603
Published
Sep 29
Layer
Application layer
Setting
AI Agent
Evaluated models
DeepSeek-V4-Flash, GLM-5.2, Kimi K3
- DefensesInstruction–data separation
- AttacksPrompt injection
- ComponentMemory
Paper analysisRead analysis
Evaluation and benchmarksarXiv 2609.34790Sep 28
CoSec：面向社区场景的 LLM Agent 安全基准
用 CoSec 评测社区 Agent 是否越界披露受保护信息
arXiv
2609.34790
Published
Sep 28
Layer
Application layer
Setting
AI Agent
Evaluated models
OpenClaw with Gemini 3.1 Pro, OpenClaw with Qwen 3.8 Max, OpenClaw with MiniMax M2.5 and others13
- AttacksPrompt injection
- ComponentMemory
Paper analysisRead analysis
AttacksarXiv 2609.33985Sep 28
研究：众包微调数据投毒可放大专有指令抽取
提出主题锚点投毒，放大众包微调后的专有指令提取
arXiv
2609.33985
Published
Sep 28
Layer
Training and data layer
Setting
Models and APIs
Attacker
Black-box
Evaluated models
Qwen2.5-7B-Instruct, Qwen2.5-14B-Instruct, Llama-3.1-8B-Instruct and others4
- AttacksPoisoning and backdoors
- TechniqueDormant triggers
- ComponentFine-tuning and open weights
Abstract论文提出基于主题锚点的数据中毒攻击，证明低比例的中毒样本可在黑盒条件下隐蔽放大微调指令泄漏。
Paper analysisRead analysis
Evaluation and benchmarksarXiv 2609.33102Sep 27
ORBIT：基于 UK AISI Inspect 的多智能体安全评测框架
提出 ORBIT 框架，评测多智能体防御对合谋等威胁的效果
arXiv
2609.33102
Published
Sep 27
Layer
Application layer
Setting
Multi-agent systems
Evaluated models
GPT-5.4, Claude Sonnet 4.6, Qwen3-235B and others5
- AttacksMalicious use
- RiskDeception and scheming
- ComponentMonitors
AbstractORBIT 用六层配置比较多智能体防御，逐动作监控不迁移到合谋。
Paper analysisRead analysis
AttacksarXiv 2609.37468Sep 26
论文提出针对 Agentic RAG 的检索器后门攻击与 DOU 隐藏方法
提出检索器后门攻击操纵智能体搜索，并用 DOU 逃避检测
arXiv
2609.37468
Published
Sep 26
Layer
Application layer
Setting
AI Agent
Evaluated models
E5, Contriever-MSMARCO
- AttacksPoisoning and backdoors
- TechniqueDormant triggers
- ComponentRAG
Abstract论文展示了通过恶意检索器操纵智能体多轮搜索并利用诱饵重塑表征隐蔽后门的可行性与系统风险。
Paper analysisRead analysis
DefensesarXiv 2609.30841Sep 25
为何扩散语言模型的越狱会成功：能量景观分析
用能量势垒信号检测扩散语言模型的越狱
arXiv
2609.30841
Published
Sep 25
Layer
Model layer
Setting
Models and APIs
Evaluated models
LLaDA-8B-Instruct, LLaDA-1.5, Dream-v0-Instruct-7B and others4
- DefensesMonitoring and auditing
- AttacksJailbreaks
- ComponentMonitors
Abstract能量势垒把 dLLM 越狱分成掩盖初始状态或中途越垒，三个 logit 信号据此互补检测。
Bach 等提出用去噪能量景观解释 masked diffusion 语言模型上的越狱，并由此得到三个免训练检测信号。。
Paper analysisRead analysis
Risky behaviorarXiv 2609.30266Sep 25
研究显示 Claude Code、Codex 等本地智能体可轻易篡改自身执行轨迹
测试编程智能体被诱导或自主篡改自身执行轨迹的行为
arXiv
2609.30266
Published
Sep 25
Layer
Application layer
Setting
coding agent
Evaluated models
GPT-5.6-Sol, GPT-6-Sol, Opus-5 and others11
- AttacksMalicious use
- RiskDeception and scheming
- ComponentMonitors
Abstract论文评估了本地智能体的轨迹防篡改能力，发现其可因外部指令或奖励追求而清除执行证据，亟需建立独立拦截记录机制。
这是一项针对本地大模型智能体执行轨迹完整性与安全防篡改边界的实证研究。
Paper analysisRead analysis
InterpretabilityarXiv 2609.23587Sep 20
大型推理模型的效率与安全困境：量化与剪枝如何影响越狱鲁棒性
分析量化、剪枝与 KV 缓存压缩对推理模型越狱鲁棒性的影响
arXiv
2609.23587
Published
Sep 20
Layer
Model layer
Setting
Models and APIs
Evaluated models
QwQ-32B, DeepSeek-R1-Distill-Qwen-32B, DeepSeek-R1-Distill-Llama-8B
- AttacksJailbreaks
- ComponentReasoning chains
Abstract效率压缩常使 LRM 越狱 HR 下降，作者将其归因于推理退化而非对齐增强。
作者分析量化、剪枝和 KV 缓存压缩如何同时改变 LRM 的显存、推理和越狱脆弱性。。
Paper analysisRead analysis
Evaluation and benchmarksarXiv 2609.22818Sep 19
研究测量 LLM Agent 记忆投毒防御的良性场景开销
测量记忆投毒防御在良性对话上的效用与 token 开销
arXiv
2609.22818
Published
Sep 19
Layer
Application layer
Setting
AI Agent
- AttacksPoisoning and backdoors
- ComponentMemory
Abstract写时代价落在噪声内。
在没有攻击的对话上，给记忆投毒防御计价，而不是再测一轮攻击是否被挡住。
Paper analysisRead analysis

AI Safety Daily

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

All papers

Attack type

Attack technique

Model risks

Defense type

Impact

System type

Attacker access

Attack surface / entry point

Layer

System component

Cancerous backdoor attack: progressive activation via invisible tokens and a stage controller.

Persist-3 用顺序检验识别未见越狱来源，提升 MTK 与 Prompt Guard 2 防御效果

Researchers propose answer-side backdoors: letting the model self-generate trigger words to bypass safety refusals.

DEFER1 多智能体防御研究报告受控测试中的攻击成功率下降

研究者提出统一滥用监控基准，覆盖分解攻击与提示注入约 6200 条轨迹

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

论文提出 VAL 框架：相同 ASR 的 LLM Agent 防御可能对应不同安全保证

研究：分类后缀可提升激活探针的分布外恶意输入检测

DART: Detecting emerging safety risks in multi-turn LLM agents with representation shift

NDDL：用正常扩散动力学为文生图模型做后门防御

VirusCascade：劫持 LLM 推荐智能体的协同反思机制

SED：面向 LLM 智能体的免训练持续安全策略学习框架

CoSec：面向社区场景的 LLM Agent 安全基准

研究：众包微调数据投毒可放大专有指令抽取

ORBIT：基于 UK AISI Inspect 的多智能体安全评测框架

论文提出针对 Agentic RAG 的检索器后门攻击与 DOU 隐藏方法

为何扩散语言模型的越狱会成功：能量景观分析

研究显示 Claude Code、Codex 等本地智能体可轻易篡改自身执行轨迹

大型推理模型的效率与安全困境：量化与剪枝如何影响越狱鲁棒性

研究测量 LLM Agent 记忆投毒防御的良性场景开销