AI Security AI安全 7h ago Updated 2h ago 更新于 2小时前 49

Russian Espionage Group Exploited Zimbra Zero-Day to Steal Mail and 2FA Codes 俄罗斯间谍组织利用Zimbra零日漏洞窃取邮件和双因素认证代码

A Russian state-sponsored espionage group (TA488) exploited a zero-day stored XSS vulnerability (CVE-2025-66376) in Zimbra Collaboration Server to compromise Western government and commercial entities. The exploit is "zero-click" regarding user interaction, executing malicious JavaScript via CSS @import tag-splitting techniques when an email is merely viewed or rendered. The malware, named ZimReaper, exfiltrates emails, browser passwords, 2FA scratch codes, and creates persistent app-specific pa 俄罗斯国家支持的黑客组织TA488利用Zimbra邮件系统的零日漏洞CVE-2025-66376,通过“查看即执行”的存储型XSS攻击窃取西方政府和企业用户的邮件及2FA验证码。 攻击载荷ZimReaper通过巧妙的HTML标签拆分技术绕过Zimbra的 sanitizer,窃取CSRF令牌、浏览器保存密码、2FA恢复码,并通过DNS查询外泄数据。 该漏洞影响Zimbra Collaboration 10.0(<10.0.18)和10.1(<10.1.13),攻击者利用生成的特定应用密码维持持久访问,即使受害者重置主密码也无法清除后门。 补丁仅修复漏洞而不撤销已窃取的凭证,受影响用户必须立即升

75
Hot 热度
65
Quality 质量
70
Impact 影响力

Analysis 深度分析

TL;DR

  • A Russian state-sponsored espionage group (TA488) exploited a zero-day stored XSS vulnerability (CVE-2025-66376) in Zimbra Collaboration Server to compromise Western government and commercial entities.
  • The exploit is "zero-click" regarding user interaction, executing malicious JavaScript via CSS @import tag-splitting techniques when an email is merely viewed or rendered.
  • The malware, named ZimReaper, exfiltrates emails, browser passwords, 2FA scratch codes, and creates persistent app-specific passwords to maintain access even after password resets.
  • Zimbra released patches for versions 10.0.18 and 10.1.13 in November 2025, but the vulnerability was actively exploited for at least five months prior to the fix.
  • Immediate remediation requires patching, auditing logs for specific SOAP calls and IMAP preferences, and resetting credentials for any affected accounts.

Why It Matters

This incident highlights the critical risk of "view-based" exploits in enterprise communication platforms, where traditional security controls like click-awareness are bypassed by automatic rendering engines. It demonstrates how sophisticated threat actors can leverage obscure sanitization flaws in widely used software to establish long-term persistence and exfiltrate high-value intelligence from critical infrastructure sectors. For AI and cybersecurity practitioners, it underscores the necessity of monitoring client-side rendering behaviors and validating third-party library updates against emerging zero-day threats.

Technical Details

  • Vulnerability Mechanism: CVE-2025-66376 is a stored cross-site scripting (XSS) flaw in Zimbra's Classic UI. Attackers use a "tag-splitting" technique, hiding an <svg onload> tag within a display:none div and breaking it apart with fake @import directives and HTML comments. The Zimbra sanitizer strips the imports, leaving behind executable JavaScript that runs upon message render.
  • Malware Capabilities (ZimReaper): The payload steals CSRF tokens, autofilled passwords, and 2FA recovery codes. It uses the platform's own APIs to brute-force the Global Address List and exfiltrates 90 days of email as a TGZ archive over DNS queries to Command and Control (C2) servers.
  • Persistence Method: The malware creates an app-specific password named ZimbraWeb via CreateAppSpecificPasswordRequest, granting IMAP/POP3/SMTP access without requiring two-factor authentication. It also sets zimbraPrefImapEnabled to TRUE, ensuring access survives password resets.
  • Target Scope: The campaign targeted government, defense, transportation, and financial organizations in NATO states, Ukraine, the CIS, Africa, and US nuclear installations. C2 infrastructure consisted of at least nine IPs and domains, each active for an average of 35.4 days.
  • Remediation Steps: Organizations must upgrade to Zimbra 10.1.20 or later, audit /opt/zimbra/log/audit.log for CreateAppSpecificPassword calls, disable IMAP for non-essential accounts, alert on GetScratchCodesRequest SOAP calls, and filter DNS for known C2 domains.

Industry Insight

  • Shift in Exploit Delivery: Security teams must adapt to threats that do not require user clicks. Email gateways and clients should prioritize strict content sanitization and disable automatic HTML rendering for untrusted sources to mitigate view-based XSS attacks.
  • Credential Hygiene and Persistence: The ability of malware to create app-specific passwords that survive main account resets indicates a need for stricter monitoring of API-level credential generation. Regular audits of active sessions and app-specific tokens are essential for detecting hidden persistence mechanisms.
  • Supply Chain and Third-Party Risk: As Zimbra is a common platform for enterprises, the delay between vulnerability discovery and patching (five months in this case) emphasizes the importance of proactive vulnerability management and rapid response protocols for critical communication infrastructure.

TL;DR

  • 俄罗斯国家支持的黑客组织TA488利用Zimbra邮件系统的零日漏洞CVE-2025-66376,通过“查看即执行”的存储型XSS攻击窃取西方政府和企业用户的邮件及2FA验证码。
  • 攻击载荷ZimReaper通过巧妙的HTML标签拆分技术绕过Zimbra的 sanitizer,窃取CSRF令牌、浏览器保存密码、2FA恢复码,并通过DNS查询外泄数据。
  • 该漏洞影响Zimbra Collaboration 10.0(<10.0.18)和10.1(<10.1.13),攻击者利用生成的特定应用密码维持持久访问,即使受害者重置主密码也无法清除后门。
  • 补丁仅修复漏洞而不撤销已窃取的凭证,受影响用户必须立即升级版本并手动重置密码、使会话失效及重新生成2FA代码。

为什么值得看

这篇文章揭示了针对关键基础设施(如核设施、国防工业)的高级持续性威胁(APT)如何利用邮件系统的安全缺陷进行大规模情报窃取,强调了“零点击”或“低交互”漏洞在间谍活动中的巨大破坏力。对于AI从业者和安全专家而言,它展示了恶意软件如何通过API滥用和数据外泄技术规避传统检测,为构建更智能的异常行为检测模型提供了实战案例。

技术解析

  • 漏洞原理与绕过技术:CVE-2025-66376是一个存储型XSS漏洞。攻击者利用CSS @import 处理机制,将 <svg onload=eval(atob(...))> 标签拆分为多个片段并隐藏在 display:none 的div中。Zimbra的HTML sanitizer无法识别这些被注释和伪指令分割的碎片,导致它们在渲染时重新组合成可执行的JavaScript代码。
  • 恶意载荷功能 (ZimReaper):Payload不仅窃取最近90天的邮件和全局地址簿,还通过平台API获取2FA刮刮卡代码和Zimbra版本信息。它使用DNS长随机子域名查找将数据外泄到C2服务器,并利用 CreateAppSpecificPasswordRequest API创建名为“ZimbraWeb”的应用程序专用密码,从而绕过2FA限制获取IMAP/POP3/SMTP访问权限。
  • CVSS评分争议:NVD将该漏洞评分为6.1(需用户交互),而MITRE评分为7.2(无需用户交互)。Unit 42指出,由于只需“查看”邮件即可触发执行,实际上属于零点击攻击,这反映了传统评分标准对现代Web应用渲染引擎安全性的评估滞后。
  • 持久化机制:攻击者不仅窃取凭据,还修改用户配置项 zimbraPrefImapEnabled 为TRUE,并创建不受密码重置影响的特定应用密码。这意味着即使管理员重置了用户密码,攻击者仍可通过应用密码继续访问邮箱,实现了真正的持久化控制。

行业启示

  • 重视邮件网关与客户端的Sanitizer安全性:Zimbra案例表明,基于规则或简单正则的HTML清洗器容易被高级混淆技术绕过。安全架构应引入更严格的沙箱渲染机制或基于DOM Clobbering的检测逻辑,防止恶意脚本在客户端执行。
  • 应用程序专用密码(App Passwords)的双刃剑效应:虽然旨在提升用户体验,但此类机制若缺乏严格的监控和异常检测(如非典型地理位置或设备登录),极易成为APT组织的后门。企业应限制此类密码的生成权限,并对相关API调用进行实时审计和告警。
  • 应急响应需超越补丁安装:修复漏洞只是第一步。鉴于凭证窃取和持久化技术的存在,组织在打补丁后必须执行全面的账户清理流程,包括重置所有密码、吊销所有活跃会话、检查并删除可疑的应用程序专用密码,以及审查日志中的异常API调用。

Disclaimer: The above content is generated by AI and is for reference only. 免责声明:以上内容由 AI 生成,仅供参考。

Security 安全