Huntress Reports 40 Infections from ChatGPT Custom GPTs Delivering RAT via ClickFix and Canon DLL Sideloading
Attackers leveraged ChatGPT Custom GPTs and ClickFix lures to infect 40 users with a feature-rich RAT via multi-stage DLL sideloading and WAV-stored payloads. Huntress technical reporting reveals the full chain and evasion tactics. This extends a documented pattern of AI platform abuse requiring platform-side behavioral controls rather than user warnings alone.
Expect expanded monitoring of sponsored search results and Custom GPT listings. Platforms will likely add behavioral detection for links to Google Sites or ClickFix patterns within generated responses. Procurement records for enterprise AI safety tooling will accelerate through year-end.
Huntress: More than 150 additional infections via Custom GPT ClickFix campaigns will be recorded by end of Q4 2026.
Sources (3)
- [1]Huntress Threat Intelligence Report(https://www.huntress.com/resources/threat-reports)
- [2]The Hacker News Coverage(https://thehackernews.com/2026/09/attackers-abuse-chatgpt-custom-gpts-to.html)
- [3]Octowave Loader Analysis(https://www.malwarebytes.com/blog/threat-analysis/2025/03/octowave-loader)