作者: wangzi

  • EvilAI as AI-enhanced Tools to Exfiltrate Sensitive Browser Data and Evade Detections

    EvilAI as AI-enhanced Tools to Exfiltrate Sensitive Browser Data and Evade Detections

    EvilAI as AI-enhanced Tools to Exfiltrate Sensitive Browser Data and Evade Detections

    A sophisticated malware campaign has emerged that leverages artificial intelligence to create deceptively legitimate applications, marking a significant evolution in cyberthreat tactics.

    The EvilAI malware family represents a new breed of threats that combines AI-generated code with traditional trojan techniques to infiltrate systems worldwide while maintaining an unprecedented level of stealth.

    The malware operates by disguising itself as productivity and AI-enhanced tools, complete with professional interfaces, valid digital signatures, and functional features that align with their advertised purposes.

    Applications such as “Recipe Lister,” “Manual Finder,” and “PDF Editor” provide genuine utility to users while simultaneously executing malicious payloads in the background.

    This dual-purpose approach significantly reduces user suspicion and allows the malware to establish persistence before detection.

    Global telemetry data reveals the campaign’s extensive reach, with infections spanning multiple continents and affecting critical sectors including manufacturing, government services, and healthcare.

    Europe has reported the highest concentration of cases with 56 incidents, followed by the Americas and AMEA regions with 29 cases each.

    The rapid geographic distribution within just one week of monitoring indicates an active and expanding threat landscape.

    Trend Micro researchers identified that EvilAI employs sophisticated social engineering tactics combined with AI-generated code that appears clean and legitimate to static analysis tools.

    The threat actors create entirely novel applications rather than mimicking existing software brands, making detection considerably more challenging for traditional security solutions.

    Advanced Infection and Persistence Mechanisms

    The malware’s infection chain begins when users launch seemingly legitimate applications, triggering a covert Node.js execution process that remains hidden from user visibility.

    EvilAI as AI-enhanced Tools to Exfiltrate Sensitive Browser Data and Evade Detections
    EvilAI’s observed infection flow (Source – Trend Micro)

    The attack leverages a carefully orchestrated command sequence that silently launches node.exe via Windows command line, executing JavaScript payloads stored in temporary directories.

    The persistence mechanism demonstrates remarkable sophistication through multiple redundant methods.

    EvilAI creates scheduled tasks named “sys_component_health_{UID}” that masquerade as legitimate Windows processes, triggering daily at 10:51 AM and repeating every four hours. The implementation uses the following command structure:

    schtasks /Create /TN "sys_component_health_{UID}" /TR "\"C:\Windows\system32\cmd[.]exe\" /c start \"\" /min \"%^LOCALAPPDATA^%\Programs\nodejs\node[.]exe\" \"%^LOCALAPPDATA^%\TEMP\{UID}or[.]js\"" /SC DAILY /ST 10:51 /RI 240 /DU 24:00 /F

    Additionally, the malware establishes registry entries in the Windows Run key, ensuring execution at user logon while creating Start Menu shortcuts to maintain the illusion of legitimate software installation.

    The JavaScript files consistently follow naming patterns with GUID suffixes ending in characters such as “or,” “ro,” or “of.”

    EvilAI’s detection evasion capabilities extend beyond traditional obfuscation through the implementation of anti-analysis loops using MurmurHash3 32-bit hashing.

    These loops create the appearance of potentially infinite execution cycles to static analysis tools while actually executing only once, effectively forcing analysts to rely on dynamic analysis methods rather than static code examination.

     

  • New Malvertising Campaign Leverages GitHub Repository to Deliver Malware

    New Malvertising Campaign Leverages GitHub Repository to Deliver Malware

    A sophisticated malvertising campaign has emerged, exploiting GitHub repositories through dangling commits to distribute malware via fake GitHub Desktop clients.

    This novel attack vector represents a significant evolution in cybercriminal tactics, leveraging the trust and legitimacy associated with GitHub’s platform to deceive unsuspecting users into downloading malicious software.

    The campaign operates by promoting compromised GitHub repositories containing dangling commits that serve as delivery mechanisms for malware payloads.

    When users search for GitHub Desktop through compromised advertisements, they are redirected to malicious repositories that appear legitimate but contain hidden malware embedded within the repository structure.

    The attack leverages users’ familiarity with GitHub’s interface and their trust in the platform’s security.

    Upon successful infection, the malware establishes persistence on victim systems while maintaining covert communication channels with command and control servers.

    Attack chain (Source – X)

    Unit 42 researchers identified this campaign through behavioral analysis of suspicious GitHub repository activities and anomalous download patterns associated with fake GitHub Desktop installers.

    Advanced Infection Mechanism and Payload Execution

    The malware employs a sophisticated multi-stage infection process that begins when users download what appears to be a legitimate GitHub Desktop installer.

    The initial payload performs comprehensive system discovery, collecting detailed information about the infected machine including operating system details, installed software, and network configurations.

    This reconnaissance data is immediately exfiltrated to attacker-controlled servers before proceeding to the next infection stage.

    The campaign demonstrates particular sophistication in its use of conditional payload deployment based on system characteristics.

    PowerShell-based payloads download NetSupport Remote Access Trojan from command and control infrastructure, while executable variants deploy AutoIT interpreters with COM file extensions to evade detection.

    The malware establishes registry-based persistence mechanisms and utilizes legitimate system utilities like MSBuild.exe and RegAsm.exe for data exfiltration, effectively blending malicious activities with normal system operations.

    Detection evasion techniques include enabling browser remote debugging capabilities, setting Windows Defender exclusion paths, and leveraging trusted system processes for payload execution, making traditional security solutions less effective against this sophisticated threat.

  • 世界,您好!

    欢迎使用 WordPress。这是您的第一篇文章。编辑或删除它,然后开始写作吧!