An Approach for a Supporting Multi-LLM System for Automated Certification Based on the German IT-Grundschutz
本文针对NIS2指令实施后德国企业面临的信息安全认证需求激增、专业人员短缺及实施成本高昂的挑战，提出了一种基于多大型语言模型系统（MLS）与混合检索增强生成（HybridRAG）的半自动化BSI IT-Grundschutz认证方法。该方法
2026-06-25 归档
今日共收录 1364 条安全情报，包含 629 个 CVE，391 篇安全通告，以及 120 篇研究论文。
← 返回归档列表Security and Privacy in Retrieval-Augmented Generation: Architectures, Threats, Defenses, and Future Directions for Building Trustworthy Systems
该综述论文系统性地探讨了检索增强生成（RAG）系统中的安全与隐私挑战。RAG通过将检索机制与生成模型相结合，提升了大型语言模型的事实准确性和领域适应性，但同时也引入了检索索引、查询日志、上下文构建和联邦更新等环节的敏感信息泄露风险，以及知识
Speculative Decoding at Temperature Zero: A Scoped Safety-Invariance Screen with a 48,072-Sample Expansion
本文针对推测解码（Speculative Decoding）在温度为零（temperature zero）时的安全不变性问题进行了系统性研究。推测解码通过让一个草稿模型（draft model）生成候选 token，再由目标模型（targe
FirmCure:Towards Autonomous and Adaptive Rehosting of Linux-Based Firmware
本文提出FirmCure，首个基于大型语言模型（LLM）的全系统固件重托管框架，旨在实现Linux固件的自主、自适应重托管。固件重托管是安全分析的基础，但现有方法高度依赖专家干预，难以处理定制化设备的初始化与运行时障碍。FirmCure通过
Poisoned Playbooks: Demystifying Knowledge Poisoning Effects on AI Security Agents
本论文研究了针对基于检索增强生成（RAG）的AI安全代理的知识投毒攻击。AI安全代理越来越多地依赖RAG从外部安全知识源（如漏洞报告、CTF题解）获取信息，以进行漏洞分析和利用推理。这种机制引入了新的风险：攻击者可以通过注入精心构造的恶意文
Securing LLM-Agent Long-Term Memory Against Poisoning: Non-Malleable, Origin-Bound Authority with Machine-Checked Guarantees
本文研究了LLM智能体长期记忆面临的关键安全漏洞：记忆投毒。攻击者可以在一个会话中存储恶意内容，影响后续会话中的关键操作（如支付、设置更改、数据窃取）。现有防御机制基于内容（检测或信任评分）或衍生历史（谱系）来判定记忆项的权限，但本文证明这
Exposing the Illusion of Erasure in Knowledge Editing for LLMs
本文研究了大型语言模型（LLM）中知识编辑（Knowledge Editing, KE）方法的可靠性与潜在安全漏洞。KE 旨在无需昂贵重训练即可更新模型中的特定事实，但最新研究发现其效果存在严重局限性。作者从对抗性诱发（adversaria
Understanding the (In)Security of Vibe-Coded Applications
本文系统研究了基于大型语言模型（LLM）的“氛围编码”（vibe coding）安全缺陷。vibe coding 是一种新兴的开发范式，用户通过自然语言与AI代理交互即可创建应用，因门槛低而快速普及。不同于传统AI辅助编程中开发者仍负责实现
Black-Box Forensics for Conversational LLM Agents
本文研究了对黑盒对话式LLM智能体进行数字取证的方法。随着LLM驱动的诈骗日益猖獗，识别隐藏端点背后的基础模型（归属）以及判断两个端点是否运行完全相同的系统提示（指纹识别），对于追溯诈骗来源、揭露犯罪网络和监控API变更至关重要。作者提出了
Decoupling Reconnaissance and Exploitation: Measuring the Capability Boundaries of LLM-Based Web Penetration Testing
大型语言模型（LLM）在自动化渗透测试中展现出潜力，但现有端到端黑盒评估容易受到错误级联的影响：早期侦察阶段的失败可能掩盖代理实际利用漏洞的能力。为了更准确地刻画这些能力，本文提出了一种两阶段解耦评估框架，将利用执行与侦察分离。通过使用真实
CrowdStrike Named a Leader in Frost & Sullivan 2026 Radar for Cloud-Native Application Protection Platforms
CrowdStrike Named a Leader in Frost & Sullivan 2026 Radar for Cloud-Native Application Protection Platforms
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
CNAPP evolution: How Microsoft aligns with leading cloud risk management platforms
Learn how CNAPP platforms are helping organizations prioritize exploitable risks, reduce exposure, and operationalize se