Hackers Exploit ChatGPT Custom GPTs in Multi-Stage Malware Attack
Hackers have discovered a way to exploit OpenAI's custom GPT feature on chatgpt.com, using it as an opening move in a multi-stage malware attack.
The attackers created malicious Custom GPTs with innocuous-sounding names and used them to steer conversations towards actions that ultimately sent victims to a Google Sites page masquerading as a Cloudflare CAPTCHA verification.
The fake CAPTCHA page prompted users to execute a PowerShell command, which kicked off an eight-stage ClickFix infection chain. The command fetched a malicious MSI installer, which deployed a remote access trojan with features including remote desktop control, audio and video capture, and the ability to drop additional payloads onto the compromised system.
The attackers used decimal notation instead of standard dotted IP format in some variants of the attack, making it harder for casual inspection or URL filtering tools to detect. The distribution method was particularly clever, with victims often landing on the malicious Custom GPTs through sponsored Google search results for 'chatgpt.'