Fake Claude AI App in Bing Ads Delivers SectopRAT Malware
Attackers exploited Bing's ad platform to promote a fraudulent Claude desktop application, abusing the legitimate Claude.ai domain to lend credibility to the campaign and bypass user suspicion. The malicious installer delivered a trojanized DLL that deployed SectopRAT, a remote access trojan capable of stealing credentials, browser data, and other sensitive information. This attack highlights how threat actors weaponize trusted brand names and legitimate infrastructure to lower user defenses. The use of paid search ads as a delivery vector makes this particularly dangerous, as users inherently trust top search results. Organizations and individuals must recognize that even familiar-looking search results and domains can serve as attack surfaces.
Tactical Insight
Immediate actions
- Block or restrict end-user ability to install software downloaded from search ad links without IT approval.
- Deploy DNS filtering and web proxies to flag or block newly registered or lookalike domains associated with popular AI brands.
- Alert users via internal communications about active malvertising campaigns targeting AI application downloads.
Long-term improvements
- Enforce application allowlisting so only approved, verified software can execute on corporate endpoints.
- Establish a vetted software catalog so employees download tools only from approved, official sources rather than search engines.
- Conduct regular security awareness training that specifically covers malvertising, typosquatting, and fake software installer tactics.
Detection measures
- Monitor endpoint telemetry for suspicious DLL side-loading activity or unexpected outbound connections from newly installed applications.
- Enable browser and endpoint logging to detect downloads of unsigned or low-reputation executables originating from ad-served URLs.
- Deploy a SIEM rule to alert on SectopRAT indicators of compromise (IOCs), including known C2 infrastructure and registry artifacts.