Cyber Risks · The Hacker News

Google Uncovers PROMPTFLUX Malware That Uses Gemini AI to Rewrite Its Code Hourly

Summary: Google's Threat Intelligence Group reported on Wednesday that it found an experimental VB Script malware called PROMPTFLUX that uses a hard-coded API key to query Gemini 1.5 Flash or later for new obfuscation code, part of a 'Thinking Robot' component that logs AI responses to %TEMP%\thinking_robot_log.txt. One variant prompts the LLM to act as an 'expert VB Script obfuscator' and rewrite the malware's entire source code every hour, while the self-modification function AttemptToUpdateSelf is commented out and Google assesses the malware is in development or testing with no means to compromise a victim network or device.

What's compelling: Researcher Marcus Hutchins argued the significance is overstated, noting the prompt assumes Gemini instinctively knows how to evade antivirus, there is no entropy check to ensure new code differs, and the function was commented out and unused - a caution against treating every AI-labeled sample as a capability shift.

Read the original at The Hacker News ↗

Topics
Detection & Response
Related
YARA-X 1.23.0 Release, (Sun, Nov 15th) Cyber Risks · Nov 15, 2026