Trigger the Straggler: Load Hijack on Mixture-of-Experts LLMs
本文研究了大语言模型（LLM）中混合专家（Mixture-of-Experts, MoE）模型在专家并行（Expert Parallelism, EP）服务场景下的供应链攻击面。在EP架构中，不同专家被分布到多个GPU上，路由器（route
2026-08-12 归档
今日共收录 1910 条安全情报，包含 1065 个 CVE，437 篇安全通告，以及 97 篇研究论文。
← 返回归档列表SLAC: Access-Driven CPU-to-GPU Side-channel Attacks via System-Level Cache on Apple Silicon
该论文针对现代异构片上系统（SoC）中CPU与GPU共享最后一级缓存（LLC）或系统级缓存（SLC）所带来的新型跨域攻击面展开研究。作者以苹果Apple Silicon为目标，首次发现GPU内存访问会在共享SLC中留下集合级（set-lev
A Gateway Architecture for Enterprise MCP Authentication: Unifying Heterogeneous Auth, Identity Delegation, and the User / Non-User Persona Problem
本文报告了企业环境中模型上下文协议（MCP）大规模落地后出现的认证治理危机及一种基于中心化网关的解决方案。MCP已成为连接LLM智能体与企业工具的事实标准接口，其采用速度极快：一年内大型组织从零迅速发展到数十个内部MCP服务器。然而这种速度
Logit-Boundary Geometric Belief Interfaces and Sparse Sheaf-Enclave Protocols: A Self-Contained Substrate for Secure Network Electronic Health Record (EHR) Interoperability
本文提出了一种面向电子健康记录（EHR）互操作性的数学与工程架构，旨在解决跨系统交换时的边界问题。研究背景是：传统系统、生成式模型、术语服务、身份系统及人工审核者各自暴露丰富的内部状态，而实际操作需要一种狭窄的共享接口，以类型化声明、有界不
STAIR: Effective Incident Response Using an End-to-End Agentic Planning Framework
本文提出 STAIR，一个面向网络攻击后系统恢复的端到端智能体规划框架。传统事件响应依赖专家编写的固定剧本（playbook），但静态流程难以适应动态变化的攻击状态、恢复目标以及执行反馈。尽管已有的基于大语言模型（LLM）的规划器和工具调用
Defending Retrieval-Augmented Intrusion Detection Against Knowledge Poisoning and Prompt Injection
检索增强生成（RAG）技术通过从向量知识库中检索语义相似的历史流量，使大语言模型能够对网络流进行分类并生成可读的入侵事件报告。然而，检索层的引入同时带来了知识投毒和提示注入等安全漏洞。本文提出 RAG-IDS，一个三层多智能体入侵检测框架，
EvoTrustRAG: Evolution-Aware Conflict Attribution and Evidence Handling for Reliable Retrieval-Augmented Generation
本文针对检索增强生成（RAG）系统中证据冲突这一根本性挑战，提出了一种新的研究问题——冲突来源归因（conflict origin attribution），并设计了无训练框架 EvoTrustRAG。在动态和对抗性环境中，同一冲突可能源自
Memoir: Learning, Verifying, and Evolving False-Positive Memories for Static Application Security Testing Tools
静态应用安全测试（SAST）工具在现代安全软件开发中不可或缺，但常产生大量误报（false positive, FP）告警，导致人工审查成本高昂并降低开发者信任。现有误报缓解方法面临两大挑战：一是不同SAST工具和漏洞类别间的差异大，难以从
A flaw was found in the multicloud-integrations component. The Application propagation controller processes the `ocm-managed-cluster` annotation from an Application Custom Resou...
CVE-2026-72526
CVE-2026-72526 是 multicloud-integrations 组件中的一个高危漏洞。该组件的 Application propagation controller 在处理 Application 自定义资源（CR）中的
A flaw was found in multicloud-integrations, a component of Red Hat Advanced Cluster Management (RHACM). This vulnerability allows an authenticated user, referred to as a tenant...
CVE-2026-70398
CVE-2026-70398 是 Red Hat Advanced Cluster Management (RHACM) 中 multicloud-integrations 组件的一个安全漏洞。该漏洞源于对 GitOpsCluster 控制
Cloud & Application Security
Cloud & Application Security
From Scanner to Stealer: Inside the trivy-action Supply Chain Compromise
From Scanner to Stealer: Inside the trivy-action Supply Chain Compromise
Next-Gen Identity Security
Next-Gen Identity Security
CrowdStrike Named an Innovation and Growth Leader in the 2026 Frost Radar™: Cloud and Application Runtime Security
CrowdStrike Named an Innovation and Growth Leader in the 2026 Frost Radar™: Cloud and Application Runtime Security
CrowdStrike Expands Identity Leadership with OpenID and IDPro
CrowdStrike Expands Identity Leadership with OpenID and IDPro
cloud architecture assessment
cloud architecture assessment
ransomware and multifaceted extortion defense
ransomware and multifaceted extortion defense
Google Cloud Next 2026Catch-up sessions on the keynotes and select sessions are now available on demand.Explore content on-demand
Google Cloud Next 2026Catch-up sessions on the keynotes and select sessions are now available on demand.Explore content
The Identity Problem Hiding in AI Agent Deployments
The Identity Problem Hiding in AI Agent Deployments
94% of Organizations Report Cloud Breaches: CrowdStrike State of CDR Survey
94% of Organizations Report Cloud Breaches: CrowdStrike State of CDR Survey