iOS 15.8.8 and iPadOS 15.8.8
iOS 15.8.8 and iPadOS 15.8.8
2026-05-30 归档
今日共收录 1277 条安全情报，包含 517 个 CVE，386 篇安全通告，以及 105 篇研究论文。
← 返回归档列表MFSA 2026-49Security Vulnerabilities fixed in Firefox for iOS 151.0
MFSA 2026-49Security Vulnerabilities fixed in Firefox for iOS 151.0
MFSA 2026-52Security Vulnerabilities fixed in Firefox for iOS 151.1
MFSA 2026-52Security Vulnerabilities fixed in Firefox for iOS 151.1
Chrome Dev for Android Update
Hi everyone! We've just released Chrome Dev 150 (150.0.7863.2) for Android. It's now available on Google Play.You can se
Chrome for Android Update
Hello Everyone! We've just released Chrome 149 (149.0.7827.48) for Android to a small percentage of users. It'll become
Chrome Beta for Android Update
Hi everyone! We've just released Chrome Beta 149 (149.0.7827.48) for Android. It's now available on Google Play.You can
Firewall for AI
Akamai 发布了 Firewall for AI，这是一款专门保护 AI 应用免受恶意攻击的安全产品。随着企业广泛部署大语言模型和生成式 AI 服务，攻击者开始利用提示注入、模型滥用、数据提取等手段破坏 AI 系统的完整性和保密性。传统
KBF: Knowledge Boundary as Fingerprint for Language Model and Black-Box API Auditing
随着大语言模型（LLM）的广泛应用，第三方转售和代理API已成为访问模型的主要中间层。然而，用户无法直接验证所声称的端点是否真正运行了预期的模型，这可能导致模型替换、混合路由等欺诈行为，威胁供应链安全。本文提出KBF（Knowledge B
Cybersecurity AI (CAI) Dataset
本文介绍 CAI Dataset，一个为期十四个月的网络安全大语言模型（LLM）轨迹数据集，基于开源 CAI 代理框架收集。该数据集源于 PentestGPT 的发现：专家操作轨迹而非基础模型能力是限制网络安全 LLM 性能的瓶颈。CAI
How Reliable Are AI Attackers Against a Fixed Vulnerable Target? A 400-Run Empirical Study of LLM Penetration Testing Consistency
本研究首次大规模实证测量了大型语言模型（LLM）作为自主渗透测试代理时攻击行为的一致性。论文设计了400次自治渗透测试运行（4个模型，每个100次），目标是一个固定蜜罐环境，包含OWASP Juice Shop和两个额外脆弱服务，并保持提示
GraphSteal: Structural Knowledge Stealing from Graph RAG via Traversal Reconstruction
本文提出了一种针对图检索增强生成（Graph RAG）系统的结构知识窃取攻击框架——GraphSteal。Graph RAG 系统通过将知识图谱集成到 LLM 的检索管道中，使模型能够利用结构化知识（实体、关系、多跳依赖）增强生成质量。然而
Blind PRNG Hijacking: An Undetectable Integrity-Preserving Attack Against LLM Watermarking
本文提出一种针对大语言模型（LLM）文本水印的新型供应链攻击——SeedHack。现有密码学水印方案（如KGW、Unigram、DipMark）的安全性均基于伪随机数生成器（PRNG）的可信假设。SeedHack在供应链层面劫持PRNG，通
Towards Cybersecurity SuperIntelligence (CSI): What's the best harness for cybersecurity?
本文旨在探索网络安全人工智能（AI）的最佳“缰绳”（harness）——即用于驱动大语言模型（LLM）执行安全任务的执行框架。当前，网络安全系统正趋向于每个智能体使用单一的迭代式shell循环作为执行脚手架（scaffold），但这些脚手架
Formie is a Craft CMS plugin for creating forms. Prior to 2.2.20 and 3.1.24, unauthenticated users could submit crafted values into Hidden fields (with Default value → Custom) t...
CVE-2026-45697
Formie 是一个 Craft CMS 的插件，用于创建表单。在版本 2.2.20 和 3.1.24 之前，未认证的攻击者可以向隐藏字段（配置为“默认值→自定义”）提交精心构造的值，这些值在提交处理过程中会被作为 Twig 模板进行评估。
cpp-httplib is a C++11 single-file header-only cross platform HTTP/HTTPS library. Prior to 0.44.0, when cpp-httplib's server parses an incoming request, it applies percent-decod...
CVE-2026-45372
cpp-httplib 是一个C++11跨平台HTTP/HTTPS库，采用单文件头文件形式。在0.44.0之前的版本中，服务器解析传入请求时，会对除Location和Referer以外的所有HTTP头字段值进行百分号解码。攻击者可以利用这一
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
CrowdStrike Expands Real-Time Cloud Detection and Response to Google Cloud
CrowdStrike Expands Real-Time Cloud Detection and Response to Google Cloud
CrowdStrike Falcon Cloud Security Delivered 264% ROI Through Unified Cloud Protection
CrowdStrike Falcon Cloud Security Delivered 264% ROI Through Unified Cloud Protection
Security & Identity
Security & Identity
SAP on Google Cloud
SAP on Google Cloud