Actor Profile

The operators behind this campaign run a human-operated phishing platform that impersonates advertising products for major AI chatbots including ChatGPT, Gemini, Claude, Perplexity, Meta Muse, and Manus. The platform is part of a broader three-pronged operation that also includes Google Ads-themed refund/payment scams and recruitment-themed sites impersonating Tesla, Louis Vuitton, Nike, and Adecco. The actors' motivation appears to be financial gain through monetizing stolen advertising accounts—either by running their own ad campaigns or selling accounts with clean spend histories on underground markets. The operation demonstrates sophistication through real-time credential harvesting, device fingerprinting, and rapid adaptation to current events (e.g., launching museads.ai within a week of Meta's Muse launch in September 2026).

TTPs (Tactics, Techniques, Procedures)

The campaign employs browser-in-the-browser (BitB) phishing attacks (T1566.002 - Phishing: Spearphishing Link) to present fake login windows with spoofed address bars displaying legitimate domains like accounts.google.com or Okta tenants. The platform performs device fingerprinting and exfiltrates data to attacker-controlled endpoints via Socket.IO (T1041 - Exfiltration Over C2 Channel). Real-time credential harvesting occurs as operators monitor victim input and dynamically present appropriate MFA challenges (T1111 - Multi-Factor Authentication Interception). Initial access is achieved through fake invitation emails impersonating trusted AI brands (T1598.003 - Phishing for Information: Spearphishing Link). The shared technology stack (Next.js, Socket.IO) and common C2 endpoints across all phishing sites indicate centralized infrastructure management. OPSEC failures include exposed source code in misconfigured public GitHub repositories.

Targets & Patterns

The campaign specifically targets agency staff, media buyers, and manager-account administrators in the Technology and Artificial Intelligence sectors. Victims are selected based on their access to advertising accounts, particularly those with clean spend histories that can be monetized. The targeting strategy leverages current interest in AI chatbot advertising capabilities, with each fake portal tailored to specific platforms (ChatGPT promises Monday Google Ads briefs, Gemini offers MCC and linked-client support, Claude gets a dedicated advertising portal, etc.). The broader phishing platform also targets employees at high-profile brands (Tesla, Louis Vuitton, Nike) and uses Google Ads-themed lures, suggesting the actors cast a wide net across multiple verticals while maintaining focus on advertising account compromise. The timing of attacks demonstrates opportunistic targeting—launching new phishing sites within days of legitimate product announcements to capitalize on user curiosity and trust.

Historical Context

This campaign aligns with an established pattern of advertising account theft documented by Mimecast in July 2026, which identified malware families including VietCredCare, DuckTail, NodeStealer, and PXA Stealer as driving "widespread commodity crime in the advertising ecosystem." The current operation represents an evolution from malware-based account theft to sophisticated human-operated phishing platforms. Island researchers also linked this activity to a separate campaign abusing Google-sponsored results to deliver NetSupport RAT via custom GPTs and shared-AI chat content, with approximately 850 paid-ad landings, 26 lookalike ChatGPT destinations, and 71 Google Ads campaign IDs observed over a three-month period ending August 2026. The exposed GitHub repositories containing earlier platform versions suggest this is an iterative operation that has been refined over time.

Defensive Recommendations

  • Deploy phishing-resistant authentication (FIDO2/WebAuthn) to prevent credential capture even when users interact with spoofed login pages, mitigating T1111 and T1566.002
  • Implement browser isolation or endpoint detection capable of identifying browser-in-the-browser (BitB) attacks by monitoring for suspicious iframe/window rendering patterns and mismatched URL contexts
  • Establish continuous monitoring and alerting for advertising account administrative changes, including new administrator additions, permission downgrades, and payment method modifications
  • Block or scrutinize Socket.IO connections to unexpected domains and monitor for device fingerprinting activity via endpoints matching patterns like '/api/send/ip'
  • Conduct security awareness training focused on AI-themed phishing lures, emphasizing verification of invitation emails and legitimate vendor communication channels before connecting business accounts