Threat actors are exploiting ChatGPT’s Custom GPT feature to distribute remote access trojans by redirecting users to malicious Google Sites pages that employ deceptive ClickFix installation lures.
Key Points
- Researchers at Huntress identified a campaign where attackers created "Plus 5.6" Custom GPTs to trick at least 40 users into downloading malicious MSI installers.
- The attack chain utilizes DLL sideloading via a legitimate Canon-signed binary to execute shellcode hidden within a .WAV audio file.
- Victims are coerced into executing malicious PowerShell commands through fake Cloudflare CAPTCHA prompts, a technique known as ClickFix.
- The deployed remote access trojan (RAT) can capture camera and microphone input, steal browser data, and bypass security software like Microsoft Defender.
- Command-and-control communication is masked using DNS-over-HTTPS, allowing the malware to evade detection by standard network monitoring tools.