2026年6月2日，美国网络安全与基础设施安全局（CISA）联合联邦调查局（FBI）、国家安全局（NSA）、能源部（DOE）、环保署（EPA）、运输安全管理局（TSA）、交通部（DOT）及农业部（USDA）发布联合公告，指出近期观察到针对美国自动油箱表（ATG）系统的恶意网络活动。ATG系统广泛应用于能源、化工、食品与农业、运输系统等行业，用于远程监控储罐参数（如液位、温度、泄漏检测）。威胁行为者通过互联网暴露的ATG系统实施入侵，利用认证绕过、硬编码凭据、操作系统命令执行、SQL注入、权限提升等多种漏洞，获取设备管理界面访问权限，进而执行任意代码、操作数据库并获取完全管理员权限。攻击者可能篡改系统属性（网络设置、产品标识、油罐容量、泵控制），导致操作故障（如罐填充水平不可见），甚至禁用系统警报，增加泄漏或继电器故障等环境或物理危害风险。目前美国政府尚未将攻击归因于特定国家或组织。公告强烈建议所有ATG所有者与运营商立即采取行动：消除公网暴露（关闭默认TCP端口8001、9001、10001等），使用防火墙、ACL或VPN限制远程访问；更改所有默认密码，实施强密码和防范钓鱼的多因素认证；联系认证服务提供商进行合规检查、软件更新和安全补丁安装。

Mandiant 发布了一份威胁情报报告，详细分析了针对 KnowledgeDeliver 产品的 ViewState 反序列化漏洞利用活动。ViewState 是 ASP.NET 中用于维护页面状态的机制，当应用程序未正确验证序列化数据时，攻击者可以通过构造恶意的 ViewState 数据触发反序列化，从而在服务器上执行任意代码。Mandiant 观察到，该漏洞已被积极利用，攻击者可能通过特制的 HTTP 请求向目标的 KnowledgeDeliver 实例发送恶意 ViewState 数据，实现远程代码执行。报告指出，此攻击活动与已知的 APT 组织有关联，但未明确归因。受影响的产品版本尚未完全披露，但 Mandiant 建议用户立即检查是否使用了易受攻击的 KnowledgeDeliver 版本，并采取应急措施。由于该漏洞利用无需身份验证，攻击面较广，可能导致数据泄露、系统控制等严重后果。Mandiant 还提供了检测指标和缓解措施，包括启用请求验证、禁用不必要的 ViewState 以及升级到最新版本。目前尚无 CVE 编号，建议用户关注厂商公告。

Ivanti Endpoint Manager Mobile（原MobileIron）存在代码注入漏洞。该漏洞允许攻击者向受影响系统注入恶意代码，可能在无需用户交互的情况下实现远程代码执行。由于缺乏官方公告和详细技术细节，目前未知具体受影响版本和攻击向量。建议用户尽快关注Ivanti官方安全更新。

根据知道创宇/Seebug漏洞平台（SSVID-92331）的公告，天融信（TOPSEC）防火墙存在一个远程利用漏洞，代号为“ELIGIBLEBACHELOR”。该漏洞允许未经身份验证的远程攻击者利用防火墙的特定协议或功能缺陷，在未授权的情况下执行任意代码或导致拒绝服务。由于公告未提供具体技术细节、影响版本或验证方法，目前无法确认漏洞的准确成因和影响范围。已知该漏洞为远程利用，可能对网络边界设备造成严重威胁，但尚未有官方补丁或详细的缓解措施发布。建议用户密切关注天融信官方安全通告，并采取临时访问控制措施（如限制管理接口、启用入侵防御规则）以降低风险。

知道创宇/Seebug 漏洞库于2026年6月3日收录了一个影响 Fortigate 防火墙的远程代码执行漏洞，代号 EGREGIOUSBLUNDER。该漏洞允许攻击者可能通过未公开的向量远程执行任意代码，从而完全控制受影响的设备。目前漏洞细节、影响版本及利用方式尚未完全公开，官方补丁状态未知。建议用户密切关注 Fortinet 官方安全公告，及时获取修复方案。

CVE-2025-5777 是 Citrix NetScaler（一款应用交付控制器和负载均衡器）中的内存泄漏漏洞。该漏洞可能允许未经身份验证的攻击者通过发送特制请求导致设备内存资源耗尽，从而造成拒绝服务（DoS）或性能下降。由于缺乏官方详细公告和严重性评级，具体触发机制和受影响版本尚未明确，但已知 Citrix 产品常被用于企业网络边界，漏洞可能影响业务连续性。根据知道创宇 / Seebug 社区提供的漏洞信息，建议管理员密切关注 Citrix 官方安全公告，并及时获取补丁。

2026年6月，微软威胁情报团队披露一起大规模npm供应链攻击活动，代号Miasma。攻击者通过入侵Red Hat旗下的@redhat-cloud-services npm包，在超过90个版本中嵌入了恶意代码。该恶意代码在开发者安装或更新这些包时触发，利用npm的preinstall脚本机制实现持久化，秘密窃取目标系统上的GitHub令牌、云平台访问密钥（如AWS、Azure等）以及本地存储的凭据。窃取的凭据被用于进一步访问受害者的代码仓库和云环境，同时攻击者利用受信任的包身份重新发布被篡改的版本，使得感染像蠕虫一样传播至其他信任该包的开发者或CI/CD流水线。该攻击主要针对使用Red Hat云服务相关npm包的开发者和企业，通过供应链上游污染向下游用户扩散。微软建议相关组织立即检查@redhat-cloud-services包的使用情况，并审查npm包的完整性。目前尚未公开归因于特定APT组织或恶意软件家族。

A security vulnerability has been detected in EIPStackGroup OpENer up to 2.3.0. Affected is the function CreateMessageRouterRequestStructure of the file cipmessagerouter.c of the component SendRRData Handler. The manipulation leads to use after free. Remote exploitation of the attack is possible. The exploit has been disclosed publicly and may be used. The project was informed of the problem early

CWE-639: Authorization Bypass Through User-Controlled Key in web services in Progress Sitefinity 15.2.x before 15.2.8441, 15.3.x before 15.3.8531, and 15.4.x before 15.4.8630 allows a remote authenticated attacker to modify account properties of other users, potentially leading to account compromise. Successful exploitation requires knowledge of values that are not generally exposed to low-privile

PLANET VDR-300NU ADSL Router - 未授权修改DNS

Acknowledgements

CISA has added two new vulnerabilities to its Known Exploited Vulnerabilities (KEV) Catalog, based on evidence of active exploitation. CVE-2022-0492 Linux Kernel Improper Authentication Vulnerability CVE-2025-48595 Android Framework Integer Overflow Vulnerability These types of vulnerabilities are a frequent attack vector for malicious cyber actors and pose significant risks to the federal enter

CVE-2026-10703 是影响 EIPStackGroup OpENer 至 2.3.0 版本的一个安全漏洞。该漏洞存在于 cipmessagerouter.c 文件的 CreateMessageRouterRequestStructure 函数中，属于 SendRRData 处理程序的一部分。由于对已释放内存的引用（Use-After-Free），攻击者可以通过网络远程利用此漏洞，导致受影响组件的机密性、完整性和可用性受损（CVSS 评分 6.3，中等风险）。该漏洞的利用代码已公开披露，增加了被攻击者利用的可能性。尽管项目维护者已通过 issue 报告获知此问题，但尚未发布官方补丁或回应。建议受影响用户立即限制 OpENer 服务的网络暴露范围，并监控厂商后续更新以获取补丁。由于该组件常用于工业以太网/IP 协议栈，在 OT/ICS 环境中部署时需格外谨慎，评估潜在风险并采取缓解措施。

Medplum before 5.1.14 contains a server-side request forgery vulnerability in the subscription worker that allows authenticated users to perform unauthorized internal network requests by creating FHIR Subscription resources with arbitrary endpoint URLs. Attackers can point subscription endpoints at internal addresses such as cloud instance metadata services, internal databases, or container orches

The ARMember Premium plugin for WordPress is vulnerable to an insecure password reset mechanism in all versions up to, and including, 7.3.1. The plugin stores a plaintext copy of the password reset key in the `arm_reset_password_key` user meta field when a user requests a password reset. This is in addition to the hashed key that WordPress core stores securely in `wp_users.user_activation_key`. Th

The ARMember Premium plugin for WordPress is vulnerable to SQL Injection via the 'sSortDir_0' parameter of the `get_private_content_data` AJAX action in all versions up to, and including, 7.3.1. This is due to insufficient sanitization of the user-supplied parameter which is concatenated directly into the ORDER BY clause of an SQL query without a whitelist check. This makes it possible for authent

Dräger Perseus A500 software versions 2.00 through 2.02 contains an improper input handling vulnerability that allows external attackers to cause a denial of service by sending specifically crafted non-Medibus-compliant data through the Medibus interface. Attackers can overload the internal processor with malformed data to trigger a warm restart, causing ventilation pressure to drop to ambient lev

Dräger Atlan A350 software versions 1.00 through 1.01 contains an improper input handling vulnerability that allows attackers to cause a denial of service by sending specifically crafted non-Medibus-compliant data through the Medibus interface. Attackers can transmit malformed data to overload the internal processor, gradually disrupting device operation over several hours and causing loss of data

OpenMed before 1.5.2 contains a remote code execution vulnerability in the PII privacy-filter model loading path. The privacy-filter dispatcher used broad substring matching on the user-supplied model_name parameter, allowing a value such as attacker/foo-privacy-filter-bar to route through a path that loads Hugging Face models with trust_remote_code=True. An unauthenticated attacker can supply a m

Spacelabs Healthcare Sentinel versions 10.5.x and higher and 11.x.x before 11.6.0 contain an unauthenticated remote code execution vulnerability through a deprecated .NET Remoting HTTP channel exposed on port 8989 that allows attackers to perform arbitrary file read and write operations by supplying valid .NET URI endpoints. Attackers can write ASPX webshells to the IIS wwwroot directory to achiev

A high security vulnerability affecting Security Center main server installations has been identified. It could allow an attacker with local OS privileges to the main server to access the Server Admin credentials. A third party hired by Genetec found the issue. There is currently no evidence of active exploitation. This vulnerability is associated with specific installation package builds rather

Bitdefender Napoca bare-metal hypervisor contains an out-of-bounds write vulnerability in the BIOS INT 0x15 / E820 memory map handler, implemented in napoca/guests/bios_handlers.c. The handler computes a destination offset into the guest RealModeMemory buffer from guest-controlled ES and EDI register values without validating that the resulting address remains within the 1MB RealModeMemory allocat

SIP signaling stack in Verizon IMS (unspecified version) implements SIP signaling without IPsec integrity protection (missing Security-Client/Security-Server headers and ESP traffic), which allows an on-path attacker to compromise confidentiality, integrity, and authenticity of VoLTE signaling via passive monitoring and active manipulation of unsecured SIP messages over the radio and core network.

transmission through 4.1.1 was found to have a clickjacking weakness in the browser-facing WebUI and RPC response paths.

CWE-284: Improper Access Control in web services in Progress Sitefinity 15.4.8623 before 15.4.8630 allows a remote unauthenticated attacker to access content that should be restricted, resulting in full compromise of confidentiality, integrity, and availability of affected installations.

CWE‑522: Insufficiently Protected Credentials in web services in Progress Sitefinity version from 14.0.7700 to 14.4.8152, and 15.0.8200 to 15.0.8234, and 15.1.8300 to 15.1.8335, 15.2.8400 to 15.2.8441, 15.3.8500 to 15.3.8531, and 15.4.8600 to 15.4.8630 allows a remote unauthenticated attacker to obtain plain-text credentials used connect to Sitefinity Insight service. Successful exploitation requi

A vulnerability has been identified in RUGGEDCOM RST2428P (6GK6242-6PA00) (All versions < V4.0). The affected applications stores sensitive information in the browser cache when an authenticated user modify specific configurations. This could allow an authenticated attacker to access sensitive data stored in the browser.

Firefox browsers

Sign in to Cloud

Sign Up for Free Cloud Tier

iOS 26.4.2 and iPadOS 26.4.2

iOS 18.7.8 and iPadOS 18.7.8

iOS 18.7.7 and iPadOS 18.7.7

iOS 26.4 and iPadOS 26.4

Background Security Improvements for iOS, iPadOS, and macOS

iOS 16.7.15 and iPadOS 16.7.15

iOS 15.8.7 and iPadOS 15.8.7

iOS 26.3 and iPadOS 26.3

iOS 18.7.5 and iPadOS 18.7.5

iOS 26.2 and iPadOS 26.2

iOS 18.7.3 and iPadOS 18.7.3

Compressor 4.11.1

iOS 18.7.2 and iPadOS 18.7.2

iOS 26.1 and iPadOS 26.1

iOS 26.0.1 and iPadOS 26.0.1

iOS 18.7.1 and iPadOS 18.7.1

iOS 26 and iPadOS 26

iOS 18.7 and iPadOS 18.7

Android Code Search

Android Devices

Secure an Android device

Android 16 QPR2

Mobile network security

iOS 26.5 and iPadOS 26.5

iOS 18.7.9 and iPadOS 18.7.9

iOS 16.7.16 and iPadOS 16.7.16

iOS 15.8.8 and iPadOS 15.8.8

MFSA 2026-49Security Vulnerabilities fixed in Firefox for iOS 151.0

MFSA 2026-52Security Vulnerabilities fixed in Firefox for iOS 151.1

MFSA 2026-53Security Vulnerabilities fixed in Firefox for iOS 151.2

Hi, everyone! We've just released Chrome 149 (149.0.7827.59) for Android. It'll become available on Google Play over the next few days. This release includes stability and performance improvements. You can see a full list of the changes in the Git log. If you find a new issue, please let us know by filing a bug.Android releases contain the same security fixes as their corresponding Desktop release

Hi everyone! We've just released Chrome Beta 149 (149.0.7827.59) for Android. It's now available on Google Play.You can see a partial list of the changes in the Git log. For details on new features, check out the Chromium blog, and for details on web platform updates, check here.If you find a new issue, please let us know by filing a bug.Chrome Release TeamGoogle Chrome

Firewall for AI

Microsoft is working to address a widespread service issue affecting the mail flow pipeline for Exchange Online customers across North America and Germany. [...]

AI can help attackers generate malware, create malicious payloads, bypass simple security checks, and convert vague malicious intent into functional code. The post The Zero-Knowledge Threat Actor and the End of Responsible Disclosure appeared first on SecurityWeek.

该论文研究了检索增强生成（RAG）增强的大型语言模型（LLM）系统面临的推理成本攻击（ICA）问题。RAG系统通过多阶段流水线动态检索并合成外部知识，虽然提升了性能，但也带来了高昂的推理成本。现有ICA攻击通常假设可以直接操纵提示，这在实践中难以实现。作者认为，通过投毒外部知识库（如互联网上的网页知识）是一种更可行且更具威胁的攻击方式。为此，他们提出了一种新的攻击范式——检索增强推理成本攻击（RA-ICA），通过向外部知识语料库注入恶意文档来大幅增加RAG系统的推理计算成本。为实现该攻击，设计了CREEP（Computational Resource Exhaustion via External Poisoning）框架，该框架利用LLM代理自动生成既在语义上与查询相关、又能导致推理阶段token消耗异常增加的恶意文档。为了提升攻击效果，还提出了一种新的强化学习算法MA-GRPO（Memory-Augmented Group Relative Policy Optimization），通过从历史最优对抗文档的动态记忆中学习来微调攻击代理。在三个真实数据集上的大量实验表明，RA-ICA能够将token消耗提升高达13.12倍，成功率超过90%，同时不损害生成答案的完整性。该研究揭示了RAG系统在推理成本方面的新安全漏洞，对部署RAG服务的组织具有重要警示意义。

本文提出了一种名为 Privately Inferred Credentials (πCreds) 的新型去中心化可验证凭证系统。现有基于零知识证明的凭证方案复杂、应用特定且局限于结构化数据上的谓词。πCreds 利用可信 LLM 推理在认证数据上生成凭证，LLM 对非结构化数据的语义推理能力大幅扩展了可证明声明的范围。同时，论文形式化了两个新威胁：源受限对抗样本问题（SCAE），刻画对手操纵认证数据以获取误导性凭证的鲁棒性；以及认证隐蔽谓词投毒问题（ACPP），刻画通过对抗性模型选择造成的隐私泄露。论文描述了 πCreds 在用户数据上的应用，以及一类新型凭证——对专有软件的性质进行认证而不泄露其源代码。原型系统支持在实时金融、健康、邮件和代码源上发行凭证，并通过真实金融数据上的产品专长凭证实证研究了 SCAE 和 ACPP 威胁。

该论文提出 FORGE 系统，旨在解决漏洞披露量远超组织评估能力、而 PoC 生成、漏洞优先级排序和检测规则工程三个研究社区相互隔绝的问题。FORGE 是一个多智能体系统，包含五个专用智能体（Intel、Generator、Planner、Exploit、Detector），按固定流水线执行：（1）从 CVE 元数据生成目标脆弱应用程序；（2）通过 LLM 主裁判按四级分类（L0 无证据到 L3 完全攻陷）进行分阶段、多次利用尝试；（3）基于 OpenTelemetry 利用轨迹生成 Sigma 和 Snort 检测规则。其核心机制是“渐进式利用深度”：深度利用产生更丰富的行为轨迹用于检测工程，而不同评分带的深度数据为优先级验证提供真实基准。层次化知识架构跨评估积累情报，将构建和利用经验迁移至后续 CVE。在 CVE-GENIE 数据集的 603 个 CVE 上评估，端到端达到 67.8% 的 L1+ 利用，每个 CVE 成本 1.50 美元，覆盖 8 种编程语言和 187 个 CWE 类型。无论 EPSS 或 CVSS 波段，利用率均保持在约 68%，表明模式级可达性与基于元数据的优先级正交。L2+ 利用生成的检测规则比 L1 衍生规则具有显著更高的跨度归一化基础性（p=0.035），93.4% 的 Snort 规则对合成良性流量产生零误报。

该论文提出了一种针对大语言模型（LLM）模型合并过程的统一鲁棒攻击框架RogueMerge。模型合并通过聚合来自未经验证的公共平台的任务向量，将多种专门能力集成到单个LLM中，这暴露了关键的供应链攻击面：攻击者提供的恶意任务向量可直接写入模型权重，从而诱发或放大多种下游威胁。然而，现有攻击方法仅针对分类器的静态算术启发式后门攻击，无法有效处理生成式LLM的三个关键挑战：（1）LLM采用自回归解码，合并引入的微小参数漂移会在token间累积并快速削弱攻击效果；（2）攻击者无法获知受害者的合并配置，导致静态攻击向量易被稀释或破坏；（3）实际威胁需要泛化到优化时未见过的攻击提示，静态向量无法充分编码。RogueMerge首次以统一原则应对这三个挑战：针对自回归生成，采用联合优化在合并后显式强制攻击成功；针对未知合并设置，将攻击注入建模为随机min-max问题并通过元学习风格模拟求解；针对异构攻击提示的泛化，采用分布鲁棒优化并在LLM规模下推导可计算的一阶泰勒近似，附带可证误差界。在四种威胁、六种合并算法和超过170个合并LLM上的实验表明，RogueMerge持续优于现有攻击，且在不同合并设置下保持稳定，并能抵抗标准防御措施。该研究揭示了模型合并过程中的安全风险，为防御研究提供了重要参考。

本文提出了一种轻量级、端到端的智能合约安全审计框架，利用开源大语言模型（LLM）参数规模仅为0.6B至4B。该框架将传统审计任务解耦为四个独立模块：漏洞检测、漏洞解释、严重性分类和修复建议生成。为在小型模型上维持高精度，研究采用了秩稳定低秩适配器（rsLoRA）、知识蒸馏技术，并设计了一种链式验证（CoVe）聚合策略，用于系统筛选和整合模型生成的多个草稿响应，最终形成高准确率的审计报告。实验表明，该轻量级流水线在漏洞检测任务上达到98.25%的准确率，在生成性解释任务上对齐分数为0.4375，持续优于现有7B至34B参数的开源代码密集LLM。广泛的消融研究验证了解耦审计过程相对于统一提示方法的优越性，并揭示了一种新颖的严重性中心偏误，为未来LLM辅助审计研究建立了关键基准。

本文研究了基于大语言模型（LLM）的自动评分（AG）系统中存在的提示注入（PI）攻击威胁。随着LLM在自然语言指令遵循和背景知识方面的能力提升，教育者可以通过仅提供自然语言评分标准来便捷地部署AG系统，并取得令人满意的评分性能。然而，这种便利性也引入了新的安全隐患：攻击者可能利用提示注入漏洞，操纵评分系统无论答案实际质量如何都给出人为高分。这种攻击行为严重威胁教育评估的公平性、可靠性和完整性。本研究系统性地探讨了PI攻击在AG场景中的有效性，并评估了现有防御策略的抵御能力。通过多组基于评分标准的实验，作者发现当前的LLM-AG系统对提示注入高度脆弱，现有防御措施不能完全消除风险。研究结果揭示了这一新兴威胁的迫切性，旨在推动开发更安全、鲁棒和可信的LLM教育系统。论文适合安全研究人员、教育技术开发者及AI系统设计者阅读。

本文提出首个能够直接处理二进制可执行文件原始字节的大型语言模型（LLM），解决了现有LLM无法理解编译代码字节表示的问题。传统恶意软件分析需要将原始字节“提升”为汇编等高级表示，但该过程计算开销大且易出错。作者设计了一种定制的字节分词器（byte tokenizer），通过词汇扩展技术使LLM能够直接处理原始字节序列。在恶意软件分析任务上，该模型在恶意软件家族分类任务中达到69%的准确率，在架构分类任务中达到98%的准确率。实验表明，在训练过程中注入领域知识（如编译器行为、文件格式结构）对提升模型性能至关重要，而直接使用通用LLM则缺乏准确性和洞察力。该模型已部署给少量分析师进行试用以收集反馈。论文的主要贡献在于：首次实现了字节原生的LLM，证明了在可执行文件原始字节上直接进行语义理解与问答的可行性，并为后续在恶意软件逆向、静态分析等场景的应用奠定了基础。

本文提出 IstGPT，一种基于大语言模型（LLM）和图神经网络的工业系统异常检测工具。工业互联网系统面临复杂的 ICS 攻击，现有工具难以实时检测传感器与执行器之间的复杂依赖关系。IstGPT 首先利用工业多模态知识（包括运行数据、技术文档和系统图），通过多阶段提示工程提取传感器-执行器依赖图；然后通过 LLM 优化迭代改进图的节点准确性、边一致性和逻辑连贯性；最后，集成了改进的图神经网络与编码器-解码器架构，通过重构误差检测异常。在 9 个数据集（2 个公共、6 个模拟和 1 个真实机器人手臂数据集）上与 12 个基线对比，IstGPT 在 F1 分数和新的时间感知指标 eTaF1 上均取得最佳结果。文章还讨论了在真实工业场景中部署的可行性。

该论文提出一种名为ANCHOR的架构无关的知识图谱构建系统，用于自动化地从网络威胁情报（CTI）报告中提取结构化知识。现有CTI平台如STIX通常将威胁情报简化为孤立的指示器（IoC），而基于本体的表示能够保留语义关系以支持结构化威胁分析。然而，现有的本体对齐CTI提取方法面临三个挑战：1）模式特定管道需要手动重新配置；2）基于提示的模式包含在大型本体（如UCO）上无法扩展；3）依赖企业级LLM API与隐私约束冲突。ANCHOR通过混合本体发现机制解决了这些问题，该机制结合了搜索和导航，能够动态探索大规模本体模式，并利用SHACL验证确保类型分配符合模式。在UCO、STIX和MALOnt模式上的实验表明，ANCHOR在本体类型化和模式合规性方面优于现有基线。此外，使用本地LLM的ANCHOR在类型化性能上接近企业级LLM，从而实现了高保真的隐私保护CTI分析。该研究主要面向CTI分析师、安全研究者和知识图谱构建领域的研究人员。

Palo Alto Unit 42 于2026年6月2日更新了关于 npm 供应链威胁格局的报告。报告指出，自 Shai Hulud 事件以来，npm 生态系统面临的攻击面持续演化，出现了可蠕虫式传播的恶意软件、针对 CI/CD 管道的持久化机制以及多阶段攻击等新型威胁。这些攻击通常利用软件包依赖关系、混淆技术以及自动化构建流程中的漏洞，在开发者环境中植入后门，进而影响下游用户。报告强调了 npm 作为关键软件供应链环节的脆弱性，并讨论了包括包签名验证、依赖审计、最小权限原则、CI/CD 安全加固在内的缓解措施。尽管未披露具体样本或 IOC，但该分析为安全团队理解当前 npm 威胁态势提供了重要参考。

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 Falcon Cloud Security Delivered 264% ROI Through Unified Cloud Protection

Security & Identity

SAP on Google Cloud

Inside Google Cloud

Google Cloud Next & Events

Google Cloud Consulting

Transform with Google Cloud

Threat IntelligenceNorth Korea-Nexus Threat Actor Compromises Widely Used Axios NPM Package in Supply Chain AttackBy Google Threat Intelligence Group • 16-minute read

Google Cloud Products

Cloud & Application Security

From Scanner to Stealer: Inside the trivy-action Supply Chain Compromise

Next-Gen Identity Security

CrowdStrike FalconID Brings Phishing-Resistant MFA to Falcon Next-Gen Identity Security

CrowdStrike Named a Leader in Identity Threat Detection and Response

More than 30 npm packages under Red Hat's '@redhat-cloud-services' namespace were compromised in a supply-chain attack that distributed a new variant of the Shai-Hulud credential-stealing malware, dubbed "Miasma." [...]

无人机开源飞控 PX4-Autopilot堆栈溢出漏洞 CVE-2025-15150

MongoDB内存泄漏 CVE-2025-14847(MongoBleed)

1Panel 代理证书验证绕过导致任意命令执行漏洞(CVE-2025-54424)

ModelContext Inspector 未授权访问漏洞(CVE-2025-49596)

llamaindex SQL 注入漏洞(CVE-2025-1750)

Screen 本地权限提升漏洞(CVE-2025-23395)

Zabbix认证后SQL注入漏洞(CVE-2024-42327)

ProjectSend认证绕过漏洞(CVE-2024-11680)

PyTorch库RPC框架反序列化RCE漏洞(CVE-2024-48063)

Grafana认证后DuckDB-SQL注入漏洞(CVE-2024-9264)

SPIP BigUp Unauthenticated RCE(CVE-2024-8517)

Rejetto HFS 远程命令执行漏洞(CVE-2024-39943)

Apache HugeGraph-Server Command Execution In Gremlin（CVE-2024-27348）

Zabbix 后台延时注入(CVE-2024-22120)

CrushFTP 认证绕过漏洞（CVE-2024-4040）

Alert for CVE-2026-21992

Alert for CVE-2025-61884

Alert for CVE-2025-61882

Alert for CVE-2024-21287

Alert for CVE-2022-21500

Alert for CVE-2021-44228

Linux 内核提权 CVE-2026-31431(copy-fail)

CVE-2026-0251 GlobalProtect App: Local Privilege Escalation Vulnerabilities (Severity: MEDIUM)

CVE-2026-0257 PAN-OS: GlobalProtect Authentication Bypass Vulnerabilities (Severity: HIGH)

【安全通告】Linux Kernel "CIFSwitch" 本地权限提升漏洞风险通告（CVE-2026-46243）

论文《Q-FE: A Quantum-Native 6G Far-Edge Architecture Securing Industrial IoT Digital Twins via CSIDH-PQC and Asynchronous Federated Learning》针对第六代（6G）无线网络中工业物联网（IIoT）数字孪生（Digital Twin, DT）面临的安全与性能挑战，提出了一种量子原生架构Q-FE。当前架构依赖集中式云或移动边缘计算（MEC）服务器执行DT计算，导致往返延迟过高（难以满足亚毫秒级延迟和10^{-7}可靠性），且使用经典公钥密码在“先收获后解密”（HNDL）威胁模型下易受量子攻击。Q-FE架构包含三个协同设计组件：(i) 微型数字孪生（μDT），与6G基站和高能力端点共置；(ii) 跨层后量子密钥交换模块，将CSIDH-512同源密钥材料直接嵌入MAC层控制帧，利用其紧凑密钥（≤64字节）避免报文分片；(iii) 异步联邦学习（AFL）协议，由MEC节点上的轻量级DAG智能合约管理，消除掉队者瓶颈并防止模型投毒和女巫攻击，同时不暴露原始数据。端到端仿真（NS-3 + PySyft）表明，Q-FE相比ML-KEM/Kyber-1024减少MAC层开销62%，保持P99.9 URLLC延迟为0.78毫秒，且全局模型收敛速度比同步联邦学习提升31%。协议复杂度分析显示每轮聚合复杂度为O(N log R)，μDT切换迁移在10^4个模拟事件中完成时间为1.9±0.3毫秒。正式威胁模型证实其抵抗量子窃听、模型投毒和女巫攻击的能力。该研究适合6G网络架构师、工业物联网安全研究人员以及后量子密码与联邦学习领域的从业者阅读。

大型语言模型（LLM）在代码生成方面表现出色，但仍然容易复现训练数据中普遍存在的微妙但关键的安全漏洞。当前的校准技术（如监督微调SFT和强化学习RL）通常在序列级别应用粗粒度优化，难以针对安全缺陷的局部特性——单个错误token可能危及整个程序。为解决此问题，本文提出了树状自我对弈（TSP）框架，将安全代码生成重新定义为细粒度的顺序决策过程。TSP构建一个决策树，模型在其中探索分支轨迹——同时生成安全的“黄金路径”和脆弱的变体。通过将代码生成视为自我对弈游戏，模型学会严格区分自身的局部错误，提供密集的、基于策略的学习信号，强制在漏洞通常出现的关键决策节点进行自我纠正。实验表明，TSP显著提升了模型可靠性。在Python安全基准测试中，TSP使CodeLlama-7B的通过率（SPR@1）达到75.8%，远超SFT（57.0%）和无结构自我对弈基线。关键的是，TSP实现了稳健的分布外泛化：模型不仅将未见过的CWE类别的漏洞减少了24.5%，还成功将C/C++中学到的安全原则迁移到Python、Go和JavaScript等多种语言。这表明TSP不仅仅是记忆补丁，而是内化了抽象的、语言无关的安全逻辑。该研究适合LLM安全方向的研究者、AI安全工程师以及关注代码生成中漏洞缓解的开发人员。

零知识证明（ZKP）允许证明者在不泄露私有数据的情况下向验证者证明计算正确性，兼具隐私与可验证性。然而，生成证明计算密集，主要涉及多项式（POLY）和椭圆曲线（EC）运算。这些工作负载给硬件加速带来两大挑战：高效支持多种大精度模乘，以及在动态切换的POLY与EC阶段间保持高利用率。现有可重构加速器仅部分解决这些问题，在精度可扩展性、算法灵活性和资源效率上仍有局限。本文提出ZK-Flex，一种灵活可扩展的软硬件协同设计框架，用于加速ZKP证明生成。软件层包含POLY和EC优化器，通过硬件和工作负载感知的算法选择减少计算量；硬件集成TCore——基于Toom-Cook的多精度核心，配备灵活片上网络和链表内存机制，在有限内存容量下提升并行性。在代表性ZKP基准测试中，ZK-Flex相比现有技术实现5至11倍加速和高达3.8倍的面积效率提升，为高性能、可重构的ZKP加速建立了新基础。

这篇论文提出了Echelon，一种面向跨组织语言模型适应的训练架构，其核心设计原则是强制设备级模型状态（包括参数、激活值、优化器状态及每设备更新）不得导出到管理边界之外，从而将隐私合规性作为系统不变量。与现有联邦学习框架不同，Echelon不假设跨站模型交换后再追加隐私机制，而是从边界出发设计训练流程。在每个边界内部，设备进行本地训练；跨边界传输的唯一载荷是安全聚合后的边界级增量以及 O(1) 的协调元数据，所有传输都通过具体的审计接口暴露。这种仅允许聚合数据跨界的约束改变了优化问题：全局层面从未见过每设备更新，却仍需在广域网延迟、异构参与、节点流失和非独立同分布数据下保持稳定。Echelon结合了缓冲半异步安全聚合、陈旧感知加权、参与窗口、近端局部目标以及漂移感知外部同步控制器。在1B参数LoRA适应实验中，跨越2个边界且预算匹配的对比（三组随机种子，2488万token）中，Echelon在固定token、固定字节、固定墙钟时间和固定同步次数预算下，达到验证损失3.887±0.010，并且在低通信基线下表现最佳或并列最佳。在OpenWebText压力测试中，Echelon在各种广域网和非独立同分布设置下保持2,139–2,176 token/s的吞吐量；与隐私对等的DiLoCo+SA基线相比，Echelon-DA在广域网延迟下缩短了达到目标的时间；在模拟200ms延迟或严重非独立同分布分区下，质量下降最多仅2.2%。这项工作为合规敏感的跨组织模型训练提供了可审计、高性能的解决方案。

不经意传输（OT）是隐私保护计算的基础密码学原语，广泛应用于安全多方计算、隐私信息检索、零知识证明、口令认证密钥交换等场景。尽管OT扩展技术显著降低了平均成本，但其依赖大量基础OT批次和预处理阶段，在传输次数较少或通信延迟、客户端计算受限的场景下效率不佳。本文提出I-(OT)^2协议，一种基于二次剩余问题的1-out-of-2基础OT协议，旨在最小化接收方计算量和交互次数。该协议特别适合客户端-服务器架构中接收方为低功耗硬件（如IoT设备）的场景。通过轻量级离线预处理，I-(OT)^2将在线计算负担几乎全部转移给发送方，在线通信仅需6条消息和4个摘要。论文提供了详细协议描述和形式化安全证明，并基于C语言实现开源概念验证，在真实IoT硬件上评估。结果显示：在128位安全强度（使用3072位RSA模数）下，接收方在线平均每OT成本在桌面平台低至2.80微秒，在IoT设备上为39.90微秒，比知名的SimplestOT协议快10倍以上。该研究为资源受限设备上的隐私计算提供了高效、实用的OT方案。

差分隐私（DP）文本合成技术旨在通过生成合成数据来解锁敏感语料库，使其可用于模型训练，同时保护隐私。然而，现有评估基准存在重大局限性：它们使用的任务通常即使不经过训练也能近乎解决，因此强基准性能并不能证明DP合成数据能够替代原始数据访问。为了填补这一评估空白，本文提出了ContinuousBench——一个持续自动更新的基准，用于衡量DP合成文本带来的能力增益。每个季度，ContinuousBench会发布一个全新的训练语料库及对应的问答（QA）数据集，这些数据被设计为：（1）没有语料库则无法解决；（2）在DP下可学习，因为测试知识由数百条独立记录支持。研究人员从训练语料库中生成DP合成数据，并通过标准化的训练和评估工具衡量能力提升。ContinuousBench包含两个赛道：Geminon，一个关于虚构生物的程序生成数据集；以及News，一个持续爬取的公共新闻文章流。实验结果表明，虽然标准基准已近乎饱和，但在ContinuousBench上，非私有合成数据能够从原始语料库中传递大量知识，而最先进的DP合成方法即使在ε=100的高隐私预算下也基本无法做到这一点。这揭示了当前DP合成文本在传递新知识和能力方面的严重不足，对依赖DP合成数据作为数据访问替代方案的研究与实践提出了挑战。

本文系统评估了脉冲神经网络（SNN）的不同配置（神经元模型与脉冲编码方案）在网络入侵检测中的性能。背景：传统深度学习模型计算需求高，难以部署到边缘设备和神经形态硬件，而SNN具有低功耗、事件驱动的特性，是潜在替代方案。但SNN的神经元模型和编码方案选择对入侵检测效果的影响尚不明确。方法：作者在snntorch框架上实现了9种神经元模型与3种脉冲编码方案（速率编码、增量编码、延迟编码）的27种组合，对NSL-KDD、KDDCup99、CIC-IDS2017、CTU-13四个基准数据集进行控制消融实验，仅对原始输入做有限预处理，并使用5个随机种子确保统计可靠性。结果：发现脉冲编码方案比神经元模型对检测质量影响更大；延迟编码优于速率编码和增量编码；LeakyParallel神经元配合延迟编码在四个数据集上平均准确率92.11%，宏F1分数0.80，假阳性率2.01%，在CIC-IDS2017和CTU-13上接近完美，且推理速度最快。结论：SNN在低延迟或资源受限场景下可作为传统入侵检测方法的有效替代。该研究为SNN在网络入侵检测中的应用提供了设计指南。

该论文提出了一个名为 memorywire 的供应商中立的线格式（wire format），用于代理内存操作。当前多种代理内存框架（如 mem0、Letta/MemGPT、Cognee、Zep/Graphiti、MemoryOS、MemTensor）各自拥有独立的 SDK、存储布局和操作词汇，导致集成工作重复、迁移成本高昂，且缺乏统一的人工审查机制。memorywire 基于 JSON-Schema 2020-12 规范，定义了五种内存操作（remember、recall、forget、merge、expire）和四种内存类型（语义、情景、程序、情感），并提供了 MemoryStore 接口、fan-out 路由器以及可选的“人在回路”（HITL）治理通道，允许在写入长期存储前进行人工审核。该格式并不旨在与 Model Context Protocol (MCP) 竞争，而是与之互补。作者给出了开源参考实现，包含五个后端适配器（sqlite-vec、mem0、Letta、Cognee、pgvector）。实验评估包括：在包含 100 个事实和 50 个查询的标注数据集上，召回率@5 达到 1.000，写入延迟 p50=37.8ms，读取延迟 p50=40.6ms；对抗融合实验表明，在 1-of-N 排名 0 注入扫描（K∈{0,5,...,50}）中，Reciprocal Rank Fusion (RRF) 始终保持召回率@5=1.000，而最大融合（max fusion）在 K≥5 时跌落至 0.500 且泄漏率达 80%；跨 16 场景的适配器一致性测试通过了 80 个用例中的 68 个，零故障。本贡献并非新的算法，而是将现有组件（如 RRF、有限状态机、短时/长时记忆整合、差异批准工作流）封装成一个协议中立、经验验证的参考实现，旨在促进代理内存操作的互操作性与安全治理。

该论文提出了一种名为SS-ZKR（空间语义零知识路由）的隐私保护路由协议，旨在解决现有代理互操作性标准（如Agent-to-Agent协议和模型上下文协议MCP）在跨组织信任边界传输时无法保护载荷内容隐私的问题。当前，尽管W3C去中心化标识符和可验证凭证提供了加密身份认证，但缺乏支持基于语义的路由且无需中间解密的技术，这在GDPR、HIPAA和MiFID II等合规敏感环境中成为硬性约束。SS-ZKR作为A2A/MCP的补充层，包含三个核心机制：机制I（盲路由）利用差分隐私语义意图向量与零知识证明，在不解密载荷的情况下验证其与模式的兼容性，从而实现路由决策；机制II（自适应载荷净化）对数值字段应用(ε,δ)-差分隐私，对文本字段采用启发式语义聚合，提供可证明的隐私保证；机制III（空间到密码策略编译器）将可视化的信任区域拓扑映射为确定性零知识访问电路，允许策略制定者以直观方式定义安全域。论文提供了形式化威胁模型、意图向量的信息泄漏分析、三个机制的伪代码，以及与传统基于TEE或同态加密路由方案的计算复杂度对比。实验分析表明，SS-ZKR能够在金融服务、医疗和国防等行业中实现异构AI代理的合规编排，同时避免专有数据暴露给路由基础设施。该工作主要贡献在于首次将零知识证明与差分隐私结合应用于多智能体系统的语义路由，并提出了可操作的空间策略编译方法。

该论文针对联邦学习中异构差分隐私（HDP-FL）环境下的隐私泄露风险展开研究。在HDP-FL中，客户端可以根据自身政策选择不同的隐私预算（ε_i），服务器在聚合时利用ε值进行加权以提高模型效用，但这种ε感知聚合机制会暴露客户端的梯度结构模式，使得诚实但好奇的服务器能够进行隐私推断攻击。论文首先展示了服务器在具备梯度去噪和代理建模能力时，可推断客户端的分布属性并跨轮次链接同一客户端的更新，攻击在现实知识约束下有效。虽然洗牌模型（Shuffle Model）可通过匿名化更新源防御此类攻击，但HDP-FL中的ε感知聚合要求客户端身份已知，与洗牌模型不兼容。为此，论文提出了IntraShuffler，一个中间件防御框架，通过将客户端按隐私预算分组，并在组内进行参数级别的洗牌，破坏梯度结构与客户端之间的关联，同时保留ε感知聚合能力。在四个数据集上的实验表明，IntraShuffler将梯度可恢复性降低超过60%，代理推断准确率从0.78降至0.33，模型效用与多种聚合规则相比基本持平。该工作为保护HDP-FL中客户端隐私提供了实用方案。

OpenCTI is an open source platform for managing cyber threat intelligence knowledge and observables. Versions prior to 7.260227.0 are vulnerable to XSS in the rendering of email-message observable body data. The content of the body field isn't appropriately sanitized when being rendered. Does require user interaction but could be exploited by someone sharing stix or any of the ingester. This could

An unauthenticated user with write access to the knowledge base can store an XSS payload in a knowledge base item. This issue affects glpi: before 11.0.7.

React Router is a router for React. In versions 7.0.0 through 7.14.x of react-router and versions 2.10.0 through 2.17.4 of @remix-run/server-runtime, certain crafted requests can consume disproportionate server resources via unbounded path expansion in the __manifest endpoint, resulting in response time degradation and/or service unavailability for end users. This affects React Router Framework Mo

React Router is a router for React. In versions 7.0.0 through 7.14.1, when using Framework Mode, a combination of steps could potentially allow unauthorized remote code execution (RCE) through external requests. This attack requires the application code to have an existing prototype pollution vulnerability, which can then be leveraged in a 2-step attack where the second step triggers unauthorized

React Router is a router for React. In versions 7.0.0 through 7.14.0 and 6.7.0 through 6.30.3, certain URLs passed to the redirect function can trigger an open redirect to an external domain due to path values starting with // being reinterpreted as protocol-relative URLs. The level of impact depends on the validation done by the application prior to returning the redirect. This does not impact ap

React Router is a router for React. In versions 7.7.0 through 7.13.1, when using React Router's unstable React Server Components (RSC) APIs, there is a potential client-side Cross-Site Scripting (XSS) vulnerability in the RSC redirect handling if redirects come from untrusted sources. This does not impact applications that are not using the unstable RSC APIs in React Router. This is patched in ver

React Router is a router for React. In versions 7.7.0 through 7.13.1, when using React Router's unstable React Server Components (RSC) APIs, there is a potential client-side Cross-Site Scripting (XSS) vulnerability in the RSC redirect handling if redirects come from untrusted sources. This does not impact applications that are not using the unstable RSC APIs in React Router. This is patched in ver

React Router is a router for React. In versions 7.5.1 through 7.13.1, when using Framework Mode with pre-rendering enabled, improper neutralization of the HTTP `Location` header value can permit Cross-Site Scripting (XSS) in the statically generated HTML files if the redirect location comes from an untrusted source. This does not impact applications using Declarative Mode (``) or Data Mode (`creat

CWE-639: Authorization Bypass Through User-Controlled Key in web services in Progress Sitefinity 15.2.x before 15.2.8441, 15.3.x before 15.3.8531, and 15.4.x before 15.4.8630 allows a remote authenticated attacker to modify account properties of other users, potentially leading to account compromise. Successful exploitation requires knowledge of values that are not generally exposed to low-privile

MLflow 3.9.0 with basic-auth (`--app-name basic-auth`) fails to enforce authorization checks for multiple Gateway API 'list' endpoints. Specifically, the `BEFORE_REQUEST_HANDLERS` dictionary in `mlflow/server/auth/__init__.py` does not include entries for `ListGatewaySecretInfos`, `ListGatewayEndpoints`, and `ListGatewayModelDefinitions`. This allows any authenticated user, regardless of their ass

A flaw has been found in UTT HiPER 1200GW up to 2.5.3-170306. This impacts the function strcpy of the file /goform/formFireWall. This manipulation of the argument Profile causes stack-based buffer overflow. Remote exploitation of the attack is possible. The exploit has been published and may be used.

In setUserDisclaimerAcknowledged of CarDevicePolicyService.java, there is a possible way to bypass the user dialog when adding an account to a managed device due to a missing permission check. This could lead to local escalation of privilege with no additional execution privileges needed. User interaction is not needed for exploitation.

An Improper Access Control vulnerability in Ivanti Neurons for ITSM (cloud and on-premises) allows a remote authenticated attacker to gain administrative access.

Nextcloud is an open source content collaboration platform. In Nextcloud Server from versions 32.0.0 to before 32.0.9, and 33.0.0 to before 33.0.3, an authentication bypass vulnerability allowed attackers with knowledge of a user's password to circumvent two-factor authentication (2FA) protections. When a user initiated login with valid credentials on a 2FA-enabled account, the system created a te

Nextcloud is an open source content collaboration platform. From versions 32.0.0 to before 32.0.9, and 33.0.0 to before 33.0.3, when a user shares a folder or file with a Nextcloud Team that includes an external member (a person added via email address who does not have a Nextcloud account), the system automatically creates a public link for that external member. This public link is not displayed

Nextcloud is an open source content collaboration platform. In Nextcloud Server from versions 32.0.0 to before 32.0.9, and 33.0.0 to before 33.0.3, with the knowledge of other users’ principal URL an attacker could possibly send a request to gain full access to their calendar. Therefore, the attacker must be an authenticated user. This is because of improper authorization controls in the backend o

Detecting CVE-2026-20929: Kerberos Authentication Relay via CNAME Abuse

Hackers are exploiting a critical privilege escalation vulnerability (CVE-2026-8206) in the Kirki plugin for WordPress to take over any user account, including those belonging to administrators. [...]

The Russian hacking group known as Gamaredon has been attributed to the continued exploitation of a WinRAR vulnerability to deliver multiple malware families aimed at data theft and propagation. Per Sekoia, the activity involves the weaponization of CVE-2025-8088, a path traversal flaw in WinRAR, to launch an HTML Application payload dubbed GammaPhish, which is then used to retrieve an

Android 兼容性测试套件（CTS）是用于确保 Android 设备兼容性的官方测试工具。本公告来自 Android 安全公告板，但未涉及任何具体漏洞（无 CVE 条目）。公告内容可能涉及 CTS 版本更新、测试用例调整或兼容性要求变更，旨在帮助设备制造商通过认证。由于缺乏漏洞细节，该公告对防御团队而言直接安全影响较低，但保持 CTS 更新可确保设备安全基线。

Android安全公告发布了兼容性定义文档（CDD），该文档定义了Android设备必须满足的兼容性要求，以确保应用生态系统的一致性和安全性。CDD涵盖了硬件、软件、安全模型、权限管理等关键领域，是厂商实现Android兼容性的重要参考。虽然本次公告未涉及具体漏洞，但CDD的更新可能包含新的安全要求和最佳实践，对于维护Android生态系统的整体安全性至关重要。

本页面是Android安全公告（Android Security Bulletin）中的参考文档部分，标题为“Tools, build, and related reference”，主要提供了与Android构建工具、编译环境及相关参考资料的说明。内容不包含任何具体的漏洞信息或安全更新，属于开发者参考性质的文件。对于防守方而言，该页面不直接涉及安全修复或威胁情报，但建议关注Android安全公告中的其他实际漏洞公告以获取修复补丁。

本公告指向最新的Android兼容性定义文档（CDD）版本更新，而非具体安全漏洞。CDD规定了Android设备必须满足的硬件和软件兼容性要求，以确保应用生态的一致性和设备互操作性。此次更新可能涉及对Android版本功能、权限模型、性能标准或安全特性的新要求。虽然不直接包含已披露的CVE漏洞，但遵循CDD有助于设备厂商维护安全基线，例如强制要求安全补丁级别、加密支持或硬件安全模块。对于防守方，了解CDD变化可辅助评估设备合规性及潜在安全影响。

Android安全公告（2026年6月）已发布，但本次更新摘要中未提供具体的CVE列表、漏洞细节或受影响组件说明。公告通常涵盖Android框架、系统组件、内核及供应商组件中的安全修复。由于缺乏详细信息，无法确定具体风险。建议用户及时关注官方公告页面以获取完整内容。

Apple 于 2026 年 6 月 3 日发布安全公告，建议用户更新其 Mac 上的软件。虽然公告中未提供具体的漏洞细节或 CVE 编号，但通常此类更新旨在修复可能被攻击者利用的安全缺陷。建议用户及时检查并安装所有可用的系统及应用程序更新，以降低潜在风险。由于缺乏详细说明，本次更新的具体安全影响尚不明确，但遵循厂商的安全建议是良好的防护实践。

Apple官方发布了针对iPhone和iPad的安全更新公告，建议用户立即将设备更新至最新版本。此次更新旨在修复可能存在的安全漏洞，以提升设备安全性。公告未提供具体漏洞细节或CVE编号，但强调了及时更新软件的重要性。

Apple 安全公告页面（标题：submit your research）发布于 2026-06-03，但未提供具体漏洞描述或技术细节。该页面可能是一般性研究提交引导，而非针对特定安全问题的修补公告。由于缺乏 CVE 信息、受影响产品列表及严重性评估，无法确定实际风险范围。建议访问官方页面以获取最新更新，或关注 Apple 后续发布的安全公告。

该公告是 Apple 官方发布的安全帮助页面，标题为“Get help with security issues”。页面提供了关于如何报告安全漏洞、获取安全更新、以及了解 Apple 安全最佳实践的资源链接。该公告未提及任何具体的 CVE 漏洞或安全补丁，而是引导用户访问 Apple 的安全支持网站（https://support.apple.com/en-us/111756）。作为防御方，应关注该页面中的官方指引，但本次公告本身不包含需要立即修复的安全缺陷。

Apple 于 2026 年 6 月 3 日发布了标题为“Background Security Improvements”的安全公告，涉及底层安全机制的增强。公告未提及具体漏洞或 CVE 编号，也未说明受影响的明确产品版本。此类改进通常旨在提升系统整体安全性、修复潜在的安全弱点或引入新的防御机制。由于缺乏具体细节，防守方应视为常规安全更新，建议尽快部署。

输入内容并非真正的安全公告，而是阿里云安全公告栏目下的公安备案信息（备案号浙公网安备 33010602009975号），未包含任何漏洞、安全事件、威胁情报或防御建议。信息源链接指向公安备案系统，与网络安全无关。因此无法生成有效漏洞摘要。

Pocsuite 是由知道创宇安全研究团队打造并长期维护的一款开源远程漏洞测试框架。该框架是知道创宇安全研究团队发展的基石，保障了团队在 Web 安全研究方面的能力领先。Pocsuite 提供了一个标准化的漏洞验证和利用平台，支持多种协议和漏洞类型，便于安全研究人员快速编写和测试漏洞验证脚本。该工具在安全社区中被广泛使用，用于渗透测试、漏洞验证和安全研究。由于该框架本身并非漏洞，而是一款安全工具，因此不存在直接的安全风险。但用户需注意从官方渠道获取并保持更新，以确保使用最新版本并避免被恶意利用。

Seebug（知道创宇旗下）是一个国内权威的漏洞参考、分享与学习的安全漏洞社区平台，收录了大量漏洞及PoC，为安全研究人员和白帽子提供漏洞信息共享与交流。本页面为平台介绍，未包含特定漏洞公告或技术细节。

此份资料并非传统安全漏洞公告，而是知道创宇（Seebug）发布的一则网站安全防护服务推广内容。其核心宣传点在于：基于十年Web安全防护经验，构建多层安全防护体系，结合云端大数据与安全CDN加速，能够高效识别和拦截DDoS攻击、DNS攻击、CC攻击等黑客恶意行为，从而保障网站安全。文中未涉及任何具体软件漏洞（CVE）、受影响的软件版本、攻击利用技术细节或修复方案，因此无法作为安全事件进行分析。从防守方视角看，该内容不具备提供漏洞风险认知或修复优先级的能力。建议相关人员将注意力转向官方漏洞库（如Seebug）或CVE中登记的实际安全公告。

输入内容为公安备案信息，标题'京公网安备 11000002002063号'，来源360CERT，但实际内容仅为备案号链接，无任何漏洞或安全公告信息。无法进行漏洞分析。

该信息为360CERT的资质证书（网出证（京）字第281号），发布于360官网（licence2.html），内容为网络安全服务机构的相关认证证明，不涉及任何安全漏洞或技术风险。

输入内容不是有效的安全公告，无法生成摘要。标题“京网文〔2020〕6051-1195号”为版权/ICP备案号，来源链接为360.cn许可页面，无漏洞技术细节、影响范围或修复建议。

该输入为ICP备案信息（京ICP证080047号[京ICP备08010314号-6]），来源360CERT，发布时间为2026年6月3日。内容不涉及任何网络安全漏洞或威胁情报，仅为网站备案许可公示。无漏洞描述、无CVE、无受影响产品信息。

LibreChat is an enhanced ChatGPT clone that supports multiple AI providers. In versions up to and including 0.8.3, the Model Context Protocol (MCP) server integration resolves ${VAR} placeholders against the server's process.env during Zod schema validation of user-supplied MCP server URLs. Any authenticated user can create a malicious MCP server configuration with a URL pointing to an attacker-co

authentik is an open-source identity provider. Prior to versions 2025.12.6, 2026.2.4, and 2026.5.1, the Source stage can be bypassed by sending an empty POST. This issue has been patched in versions 2025.12.6, 2026.2.4, and 2026.5.1.

authentik is an open-source identity provider. Prior to versions 2025.12.5 and 2026.2.3, due to the implementation of stages in the SFE (Simple Flow Executor) in order to make the interface more compatible with legacy browsers, it was possible to use an XSS exploit in the AutosubmitStage. This issue has been patched in versions 2025.12.5 and 2026.2.3.

The ARMember Premium plugin for WordPress is vulnerable to an insecure password reset mechanism in all versions up to, and including, 7.3.1. The plugin stores a plaintext copy of the password reset key in the `arm_reset_password_key` user meta field when a user requests a password reset. This is in addition to the hashed key that WordPress core stores securely in `wp_users.user_activation_key`. Th

Spacelabs Healthcare Sentinel versions 10.5.x and higher and 11.x.x before 11.6.0 contain an unauthenticated remote code execution vulnerability through a deprecated .NET Remoting HTTP channel exposed on port 8989 that allows attackers to perform arbitrary file read and write operations by supplying valid .NET URI endpoints. Attackers can write ASPX webshells to the IIS wwwroot directory to achiev

OpenMed before 1.5.2 contains a remote code execution vulnerability in the PII privacy-filter model loading path. The privacy-filter dispatcher used broad substring matching on the user-supplied model_name parameter, allowing a value such as attacker/foo-privacy-filter-bar to route through a path that loads Hugging Face models with trust_remote_code=True. An unauthenticated attacker can supply a m

SIP signaling stack in Verizon IMS (unspecified version) implements SIP signaling without IPsec integrity protection (missing Security-Client/Security-Server headers and ESP traffic), which allows an on-path attacker to compromise confidentiality, integrity, and authenticity of VoLTE signaling via passive monitoring and active manipulation of unsecured SIP messages over the radio and core network.

CWE‑522: Insufficiently Protected Credentials in web services in Progress Sitefinity version from 14.0.7700 to 14.4.8152, and 15.0.8200 to 15.0.8234, and 15.1.8300 to 15.1.8335, 15.2.8400 to 15.2.8441, 15.3.8500 to 15.3.8531, and 15.4.8600 to 15.4.8630 allows a remote unauthenticated attacker to obtain plain-text credentials used connect to Sitefinity Insight service. Successful exploitation requi

CWE-284: Improper Access Control in web services in Progress Sitefinity 15.4.8623 before 15.4.8630 allows a remote unauthenticated attacker to access content that should be restricted, resulting in full compromise of confidentiality, integrity, and availability of affected installations.

Improper Neutralization of Special Elements used in an SQL Command ('SQL Injection') vulnerability in Ahmad WP Job Portal allows Blind SQL Injection. This issue affects WP Job Portal: from n/a through 2.5.1.

Incorrect Privilege Assignment vulnerability in Themeisle Masteriyo LMS PRO allows Privilege Escalation. This issue affects Masteriyo LMS PRO: from n/a through 2.20.0.

The Kirki – Freeform Page Builder, Website Builder & Customizer plugin for WordPress is vulnerable to privilege escalation via account takeover in all versions 6.0.0 to 6.0.6. This is due to the plugin accepting an arbitrary email address when a username is used in the password reset request. This makes it possible for unauthenticated attackers to send a password reset link for any user registered

Langroid is a framework for building large-language-model-powered applications. Prior to version 0.63.0, SQLChatAgent executes SQL produced by an LLM, which is influenceable by prompt injection. When configured with a database role that has privileges enabling code execution or filesystem access (e.g., PostgreSQL pg_execute_server_program, MySQL FILE, MSSQL xp_cmdshell), an attacker who can shape

Cloud Foundry UAA versions v76.12.0 through v78.12.0 are vulnerable to a private key exposure. The server contains a vulnerability where EC (Elliptic Curve) private keys are inadvertently exposed through the public /token_keys endpoint. This endpoint is designed to provide public key material for JWT token verification but incorrectly exposes private key components for EC keys. The vulnerability a

Arm Whois 3.11 contains a stack-based buffer overflow vulnerability that allows remote attackers to execute arbitrary code by supplying oversized input to the IP address or domain field. Attackers can craft malicious input exceeding 658 bytes with shellcode to overwrite the structured exception handler and gain command execution when the application processes the input.

IBM WebSphere Application Server 9.0, and 8.5 is vulnerable to potential remote code execution due to deserialization of untrusted data via JAX-WS endpoints with WS-Security.

IBM WebSphere Application Server 9.0, and 8.5 is vulnerable to remote code execution caused by the bypass of security controls.

IBM WebSphere Application Server 9.0, and 8.5 is vulnerable to identity spoofing.

A flaw has been found in dask up to 3.0. Affected by this issue is the function nunique_approx of the file dask/dataframe/hyperloglog.py of the component HLL Handler. This manipulation causes resource consumption. The attack is possible to be carried out remotely. A high degree of complexity is needed for the attack. The exploitation is known to be difficult. The pull request to fix this issue awa

A vulnerability was detected in SourceCodester Pizzafy E-Commerce System 1.0. Affected by this vulnerability is the function Login of the file /admin/admin_class_novo.php of the component Administrative Control Panel. The manipulation of the argument Username results in sql injection. The attack can be executed remotely. The exploit is now public and may be used.

Cpanel::JSON::XS versions before 4.41 for Perl allow type confusion via duplicate object keys when dupkeys_as_arrayref is enabled. decode_hv() collapses duplicate object keys into an array reference under dupkeys_as_arrayref. The branch reached for a duplicate key tests `SvTYPE (old_value) != SVt_RV && SvTYPE (SvRV (old_value)) != SVt_PVAV`, which evaluates SvRV(old_value) before establishing tha

A security vulnerability has been detected in SourceCodester Online Boat Reservation System 1.0. Affected by this vulnerability is an unknown functionality of the component Administrative Endpoint. The manipulation leads to improper authorization. The attack can be initiated remotely. The exploit has been disclosed publicly and may be used. Multiple endpoints are affected.

A vulnerability was detected in SourceCodester Online Food Ordering System 2.0. Affected by this issue is the function include of the file /index.php. The manipulation of the argument page results in file inclusion. The attack can be launched remotely. The exploit is now public and may be used.

Cpanel::JSON::XS versions before 4.41 for Perl allow denial of service via UTF-8 BOM prefixed input when a decode filter callback throws. To skip a leading 3-byte UTF-8 BOM, decode_json() advances the input scalar's string pointer past the mark with SvPV_set() and restores it only on the normal return path. When decoding aborts through a Perl exception, for example a filter_json_object callback t

A security flaw has been discovered in wonderwhy-er DesktopCommanderMCP up to 0.2.38. This impacts an unknown function of the file src/search-manager.ts of the component start_search. Performing a manipulation of the argument SearchResult[] results in inefficient regular expression complexity. It is possible to initiate the attack remotely. The exploit has been released to the public and may be us

The EmergencyWP – Dead Man's switch & legacy deliverance plugin for WordPress is vulnerable to Cross-Site Request Forgery in all versions up to, and including, 1.4.2. This is due to missing or incorrect nonce validation on the form_settings_ui (settings save handler, procedural include scope) function. This makes it possible for unauthenticated attackers to modify plugin settings including the min

A weakness has been identified in johnhuang316 code-index-mcp up to 2.14.0. Affected is the function is_safe_regex_pattern of the component search_code_advanced. Executing a manipulation of the argument regex can lead to inefficient regular expression complexity. It is possible to launch the attack remotely. The exploit has been made available to the public and could be used for attacks. Upgrading

A vulnerability was identified in wonderwhy-er DesktopCommanderMCP 0.2.37. This affects the function readFileFromUrl of the file src/tools/filesystem.ts of the component read_file. Such manipulation of the argument url leads to server-side request forgery. The attack may be performed from remote. The exploit is publicly available and might be used. The name of the patch is 53699bebba9950047bca16ac

The Passeum Ticketing plugin for WordPress is vulnerable to Stored Cross-Site Scripting in all versions up to, and including, 1.0. This is due to the `get_shop_url()` method returning the `shop_name` setting value without sanitization when it begins with "http", combined with insufficient validation in the `validate_shop_name()` function which only checks for empty values and string type. This mak

QloApps through 1.7.0, fixed in commit 64e9722, contains a weak cryptographic algorithm vulnerability that allows attackers to compromise user credentials by exploiting the use of MD5 for password hashing in the Tools::encrypt() function within classes/Tools.php, which concatenates a static cookie key with the supplied password. Attackers can perform offline brute-force attacks against the MD5 has

Out of bounds write in openSeaChest’s Trim/Unmap operation in Seagate’s openSeaChest v26.03.0 on all supported platforms allows for writing extra memory describing a range of LBAs to deallocate 16 bytes outside of the allocated space when running this operation.

Fixed a VM panic caused by unbounded recursion in the grpcfuse kernel module when a container created deeply nested directories on a bind-mounted host folder and triggered a dentry invalidation event. This issue has been fixed in Docker Desktop 4.76.0.

Decoding a maliciously-crafted MIME header containing many invalid encoded-words can consume excessive CPU.

When returning errors, functions in the net/textproto package would include its input as part of the error. This might allow an attacker to inject misleading content to errors that are printed or logged.

Out of bounds write and reads in openSeaChest’s --showSCSIDefects in Seagate’s openSeaChest v25.05.3 on all supported platforms allows for writing defect information out of bounds for very large defects lists via a very bad drive with lots of defects or a maliciously crafted SCSI device’s defect response length.

A vulnerability was determined in ahujasid blender-mcp up to 7636d13bded82eca58eb93c3f4cd8708dfdfbe8b. The impacted element is the function execute_blender_code of the file /src/blender_mcp/server.py. This manipulation of the argument code causes code injection. The attack is possible to be carried out remotely. The exploit has been publicly disclosed and may be utilized. This product adopts a rol

Out of bounds write in openSeaChest’s --showSupportedFormats in Seagate’s openSeaChest v25.05.3 on all supported platforms allows for writing 1 extra byte outside of allocated memory which sets a value to 1 via a maliciously crafted NVMe device with a bogus value in the namespace FLBAS byte.

A vulnerability was found in ahujasid blender-mcp up to 7636d13bded82eca58eb93c3f4cd8708dfdfbe8b. The affected element is the function requests.get of the file src/blender_mcp/server.py of the component ZIP File Handler. The manipulation of the argument zip_file_url results in server-side request forgery. The attack can be executed remotely. The exploit has been made public and could be used. This

(*x509.Certificate).VerifyHostname previously called matchHostnames in a loop over all DNS Subject Alternative Name (SAN) entries. This caused strings.Split(host, ".") to execute repeatedly on the same input hostname. With a large DNS SAN list, verification costs scaled quadratically based on the number of SAN entries multiplied by the hostname's label count. Because x509.Verify validates hostname

Dräger Zeus Infinity Empowered (Zeus IE) and Zeus RS C500 anesthesia workstations contain a local security vulnerability that allows unauthorized individuals with physical access to compromise software integrity via USB interface manipulation. Attackers can exploit the unprotected USB interfaces to impair therapy functions, manipulate device-processed data, or leverage the device as a pivot point

A vulnerability has been found in ahujasid blender-mcp up to 7636d13bded82eca58eb93c3f4cd8708dfdfbe8b. Impacted is the function Open of the file src/blender_mcp/server.py. The manipulation of the argument input_image_url leads to injection. Remote exploitation of the attack is possible. The exploit has been disclosed to the public and may be used. This product follows a rolling release approach fo

Dräger Core 1.0.5 and Dräger M540 Converter Service 1.0.9 contain a denial of service vulnerability that allows network-adjacent attackers to trigger high CPU load by sending specially crafted, unencrypted SDC messages during the discovery process. Attackers with access to the hospital network can send malformed SDC packets to exhaust CPU resources in the affected process, causing further SDC mess

A flaw has been found in warmcat libwebsockets up to 4.5.8. This issue affects the function lws_ssh_parse_plaintext of the file plugins/protocol_lws_ssh_base/sshd.c of the component SSH Protocol Handler. Executing a manipulation of the argument msg_len can lead to resource consumption. The attack may be launched remotely. The exploit has been published and may be used. This patch is called 3f9f0c6

Dräger Protector Software prior to version 6.4.2 contains a local privilege escalation vulnerability due to insecure file system permissions that allows local attackers to execute arbitrary code with elevated privileges. Attackers can replace binaries or loaded modules on the host system to execute code with NT SYSTEM privileges.

Dräger Protector Software prior to version 6.4.2 contains a local privilege escalation vulnerability due to insecure file system permissions that allows local attackers to execute arbitrary code with elevated privileges. Attackers can replace binaries or loaded modules on the host system to execute code with NT SYSTEM privileges.

Dräger Infinity Acute Care System and Standalone Infinity M540 patient monitors versions VG4.1.1, VG4.0.3, and lower (with VG4.2 partially affected) contain a network message handling vulnerability that allows remote attackers to inject spoofed or tampered data and cause denial-of-service conditions. Attackers can compromise network communications to modify device settings such as alarm states or

A flaw has been found in code-projects Student Admission System 1.0. Affected is an unknown function of the file /index.php. This manipulation of the argument eid/did causes sql injection. The attack is possible to be carried out remotely. The exploit has been published and may be used.

A vulnerability has been found in SourceCodester Human Resource Management 1.0. Affected by this vulnerability is an unknown functionality of the file /detailview.php of the component Employee View Page. Such manipulation of the argument employeeid leads to improper control of resource identifiers. The attack may be performed from remote. The exploit has been disclosed to the public and may be use

FastNetMon Community Edition through 1.2.9 contains an out-of-bounds read in the IPv4 packet parser. In src/simple_packet_parser_ng.cpp, after validating that the packet contains at least sizeof(ipv4_header_t) bytes (20 bytes), the code advances the local_pointer by '4 * ipv4_header->get_ihl()' (line 164) without validating that (a) IHL >= 5 (the minimum valid value per RFC 791), or (b) 4 * IHL by

The ARMember Premium plugin for WordPress is vulnerable to SQL Injection via the 'order' parameter of the 'arm_directory_paging_action' AJAX action in all versions up to, and including, 7.3.1. This is due to insufficient escaping on the user-supplied 'order' and 'orderby' parameters and the lack of sufficient preparation on the existing SQL query in the `arm_get_directory_members()` function. This