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

Oct 12 · Monday

All papers

Found 58 papers

Filter5

Attack type

Jailbreaks5 Prompt injection29 Poisoning and backdoors7 Extraction and theft8 Detection evasion11

Attack technique

Gradient and representation optimization3 Automated search6 Encoding and obfuscation5 Role-play and persuasion6 Dormant triggers3

Defense type

Detection and filtering12 Instruction–data separation4 Permissions and approval3

Impact

Harmful content8 Unauthorized actions5 Data and privacy leaks10 Output manipulation32

System component

Memory4 RAG15 Guardrails and judges19 Embeddings4

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：LLM applications System type：Multi-agent systems Attack surface / entry point：User input Attack surface / entry point：Documents and messages Layer：Application layer Clear all

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

AttacksarXiv 2610.12292Oct 9
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
Oct 9
Layer
Application layer
Setting
LLM applications
Attacker
Black-box
Evaluated models
LAYA-TD, LAYA-EN, LAYA-ML and others5
- AttacksDetection evasion
- ComponentGuardrails and judges
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
DefensesarXiv 2610.11634Oct 8
LTBD：用可学习信任边界分隔符防御提示注入
提出 LTBD，用可学习分隔符隔离可信指令与不可信数据
arXiv
2610.11634
Published
Oct 8
Layer
Application layer
Setting
LLM applications
Evaluated models
Llama3-8B-Instruct, Llama3.1-8B-Instruct, Falcon3-7B-Instruct and others4
- DefensesInstruction–data separation
- AttacksPrompt injection
AbstractLTBD 只训练四个分隔符嵌入。
LTBD 是一种不改基座权重的提示注入防御：只学习用来标出可信指令与不可信数据边界的四个 token 嵌入。
Paper analysisRead analysis
AttacksarXiv 2610.10742Oct 8
BRANCH 方法绕过 6 套多扫描器护栏系统
提出 BRANCH 方法，用分支搜索绕过多扫描器护栏
arXiv
2610.10742
Published
Oct 8
Layer
Application layer
Setting
LLM applications
Attacker
Black-box
Evaluated models
Meta Llama Guard 3 8B, Meta Llama Guard 7B, Meta Llama Guard 3 1B and others18
- AttacksDetection evasion
- TechniqueAutomated search
- ComponentGuardrails and judges
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
DefensesarXiv 2610.07276Oct 6
SAFESHIELD：小语言模型部署时安全的决策组织框架
提出 SAFESHIELD，组织小语言模型部署时的安全决策
arXiv
2610.07276
Published
Oct 6
Layer
Application layer
Setting
LLM applications
Evaluated models
Qwen3-4B-Instruct
- AttacksJailbreaks
- RiskHallucination and fabrication
- ComponentRAG
AbstractSAFESHIELD 用可审计决策轨迹组织 SLM 部署时的四项安全决策，并报告协调依赖影响端到端结果。
SAFESHIELD 是面向小语言模型的部署时安全框架，把运行时安全从单项审核改写成可检查的决策过程。
Paper analysisRead analysis
AttacksarXiv 2610.04985Oct 4
研究揭示 Jev 作为应用决策层的安全与隐私风险
评估应用决策层的输入注入与隐私推断风险
arXiv
2610.04985
Published
Oct 4
Layer
Application layer
Setting
LLM applications
Attacker
Black-box, Gray-box
Evaluated models
Jev, NanoJev
- AttacksPrompt injection
- TechniqueRole-play and persuasion
AbstractJev 的类型化决策可被输入与训练操纵，也会泄露信息，并同时支持检测与滥用。
Jev 是 TypeSafe 的 System One 模型，把应用 state 和固定问题变成 Choice、Noul 或 Score 的类型化决策与概率。作者用官方 API 和可改训练的 NanoJev，考察它作为决策层时的操纵、泄露与双重用途。
Paper analysisRead analysis
DefensesarXiv 2610.04412Oct 3
论文研究 LLM 数字孪生参与增强民主时的提示注入脆弱性
提出检测、结构化与 GSPO 流程，降低共识生成的提示注入错位
arXiv
2610.04412
Published
Oct 3
Layer
Application layer
Setting
LLM applications, Multi-agent systems
Evaluated models
LLaMA-2 7B, LLaMA-3 8B, GPT 3.5 Turbo and others14
- DefensesDetection and filtering
- AttacksPrompt injection
- ComponentGuardrails and judges
Abstract三章依次做个体偏好预测、政党知识图谱审议，以及共识生成对提示注入的脆弱性与一条防御流程。
这是一篇把 LLM 数字孪生放进增强民主的博士论文，分个体表示、政党聚合和共识生成的提示注入三章。
Paper analysisRead analysis
DefensesarXiv 2610.02432Oct 2
论文提出 LLM 输入扰动鲁棒性度量，并发布 ASA 攻击与 AttestMCP 防护
提出 ASA 与 AttestMCP，度量并降低 LLM 对输入序列扰动的敏感
arXiv
2610.02432
Published
Oct 2
Layer
Application layer
Setting
LLM applications, AI Agent
Attacker
Black-box
Evaluated models
Gemma-3-27B-Instruct, Gemma-3-4B-Instruct, Llama-3.2-3B-Instruct and others12
- DefensesDetection and filtering
- AttacksDetection evasion
- ComponentGuardrails and judges
Paper analysisRead analysis
AttacksarXiv 2610.01871Oct 1
研究者提出 IMMRAG 攻击，可从多模态 RAG 数据存储中提取医学与文档图像
提出 imMRAG 攻击，从多模态 RAG 私有图像库提取内容
arXiv
2610.01871
Published
Oct 1
Layer
Application layer
Setting
LLM applications
Attacker
Black-box
Evaluated models
Lumina-mGPT, Gemini 2.5 Flash Image Preview, CLIP ViT-B/16 and others5
- AttacksExtraction and theft
- TechniqueCross-modal payloads
- ComponentRAG
Paper analysisRead analysis
Evaluation and benchmarksarXiv 2609.39146Sep 30
MADBench：多智能体辩论安全性评测基准发布
提出 MADBench，评测多智能体辩论是吸收还是放大对抗影响
arXiv
2609.39146
Published
Sep 30
Layer
Application layer
Setting
Multi-agent systems
Evaluated models
gpt-4o, GPT-5, GPT-6 Astra
- AttacksPrompt injection
- ComponentRAG
AbstractMADBench 显示辩论能吸收部分准确率攻击，同时放大未授权读写。
MADBench 是一个评测多智能体辩论安全性的基准，用来看攻击是被辩论吸收还是被放大。
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 2609.36326Sep 29
PILLAR：面向增强检索的私有倒排索引词法查找
提出 PILLAR，用私有词法预筛与客户端重排保护 RAG 查询
arXiv
2609.36326
Published
Sep 29
Layer
Application layer
Setting
LLM applications
- ComponentRAG
AbstractPILLAR 用固定模式的私有 BM25 预筛加本地嵌入重排，在两基准上换取延迟或质量。
PILLAR 是面向公开语料的隐私 RAG：半诚实服务器学不到查询词项和访问模式，客户端仍拿到用于增强生成的 k 篇文档。
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 2609.29130Sep 24
ClaimMirage：域名中的自我声明如何改变 LLM 威胁判断
提出 ClaimMirage，以域名自述声称改变 LLM 威胁告警判断
arXiv
2609.29130
Published
Sep 24
Layer
Application layer
Setting
LLM applications
Evaluated models
Llama 3.3 70B Instruct-Turbo, Gemma 3 4B it, Gemma 3 12B it and others5
- AttacksDetection evasion
- ComponentGuardrails and judges
Abstract作者称自述声称可增减告警。
Paper analysisRead analysis
Evaluation and benchmarksarXiv 2609.24662Sep 21
DUMA-Bench：评估 LLM Agent 安全的双控多智能体基准
提出双控多智能体基准 DUMA-Bench，测量对抗环境下的策略违反
arXiv
2609.24662
Published
Sep 21
Layer
Application layer
Setting
Multi-agent systems
Evaluated models
GPT-5, GPT-5-mini, GPT-5-nano and others14
- AttacksPrompt injection
AbstractDUMA-Bench 用双控交互评测智能体安全。
DUMA-Bench 把 τ²-bench 的双控交互做成智能体安全评测，用来看策略违反如何随用户、模型和环境一起出现。
Paper analysisRead analysis
DefensesarXiv 2609.22949Sep 19
多智能体系统提示注入威胁模型与防御架构：14 类攻击向量与四类防御
提出多智能体系统提示注入的威胁模型与四层架构防御
arXiv
2609.22949
Published
Sep 19
Layer
Application layer
Setting
Multi-agent systems
Evaluated models
GPT-4o (2024-08-06), Claude 3.5 Sonnet, Llama 3 70B-Instruct
- DefensesInstruction–data separation
- AttacksPrompt injection
- ComponentPermissions and sandboxing
Paper analysisRead analysis
DefensesarXiv 2609.21088Sep 18
Provenance-Aware Transformers：用来源感知架构分离可信边界以防御间接提示注入
提出 Provenance-Aware Transformers，隔离不可信来源防御间接提示注入
arXiv
2609.21088
Published
Sep 18
Layer
Application layer
Setting
LLM applications
Evaluated models
SmolLM2-360M-Instruct, LLaMA-3-8B-Instruct, Qwen2.5-7B-Instruct and others4
- DefensesInstruction–data separation
- AttacksPrompt injection
Paper analysisRead analysis
AttacksarXiv 2609.19722Sep 17
ALIBI：向二进制注入虚假安全叙事可让 LLM 恶意软件分析器误判
提出 ALIBI，向恶意二进制注入不执行掩护叙事使分析器误判为良性
arXiv
2609.19722
Published
Sep 17
Layer
Application layer
Setting
LLM applications
Attacker
Black-box
Evaluated models
Gemini 2.5 Pro, GPT-5.5 Pro, Claude Opus 4.7
- AttacksDetection evasion
- TechniqueRole-play and persuasion
AbstractALIBI 用二进制内的掩护叙事重释可疑静态证据，Gemini 上翻转多，另两模型主要为降级。
ALIBI 是针对 LLM 恶意软件分诊的语义掩护叙事攻击：不覆盖系统指令，而把仍可见的可疑静态证据重释成良性安全产品的预期行为。
Paper analysisRead analysis
Evaluation and benchmarksarXiv 2609.17320Sep 15
Emergence World：对长时程多智能体系统开展对抗压力测试
提出 Emergence World，对长时程多智能体系统开展对抗压力测试
arXiv
2609.17320
Published
Sep 15
Layer
Application layer
Setting
Multi-agent systems
Evaluated models
Claude Opus 4.8, DeepSeek v4 Pro, Gemini 3.5 Flash and others9
- AttacksPrompt injection
- RiskMisalignment and value divergence
- ComponentMemory
Abstract作者通过长期平行世界实验指出模型对齐不具备可组合性，单模型群体会产生社会奉承、协同退缩及语言不透明等失效。
Paper analysisRead analysis
Evaluation and benchmarksarXiv 2609.15017Sep 14
PIDS-Bench：在过度防御、混淆与分布偏移下评测提示注入检测器
提出 PIDS-Bench，评测提示注入检测器在过度防御与分布偏移下的表现
arXiv
2609.15017
Published
Sep 14
Layer
Application layer
Setting
LLM applications
Evaluated models
TF-IDF + logistic regression, DistilBERT, DeBERTa-v3-FT and others7
- DefensesDetection and filtering
- AttacksPrompt injection
- ComponentGuardrails and judges
AbstractPIDS-Bench 显示高分布内 F1 仍会过度拦截外部来源的安全相关良性输入。
PIDS-Bench 是一个冻结的检测器级基准，用来在固定阈值下同时看提示注入检测和良性误报。
Paper analysisRead analysis

AI Safety Daily

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

All papers

Attack type

Attack technique

Defense type

Impact

System type

Attacker access

Attack surface / entry point

Layer

System component