Google researchers uncover criminal zero-day exploit likely built with AI
Google links a likely LLM-built criminal zero-day for an open-source admin tool to planned mass exploitation and maps AI-assisted threats.
Google Threat Intelligence Group linked a zero-day exploit for a popular open-source web-based administration tool, enabling 2FA bypass with valid credentials via a semantic logic error, to a criminal group, citing educational docstrings, a hallucinated CVSS score, and textbook Python as signs of LLM authorship; the vendor was notified before a planned mass exploitation campaign. The report also details Russia-nexus malware families CANFAIL and LONGSTREAM using AI-generated decoy code, the PROMPTSPY Android backdoor driving the UI through the Gemini API, APT27 using Gemini to build relay tooling, and the TeamPCP (UNC6780) supply chain compromise of LiteLLM and Trivy repositories that planted the SANDCLOCK credential stealer.
- Zero-day 2FA bypass stemmed from a hardcoded trust assumption contradicting authentication enforcement
- CANFAIL and LONGSTREAM hide malicious logic in LLM-authored filler code
- PROMPTSPY sends UI layouts to the Gemini API to autonomously tap, swipe, and capture PINs
- TeamPCP embedded SANDCLOCK in LiteLLM and Trivy builds, stealing AWS keys and GitHub tokens
- ATT&CK
- T1556 Modify Authentication ProcessT1195.002 Compromise Software Supply ChainT1528 Steal Application Access TokenT1027 Obfuscated Files or Information
- Organizations
- Google Threat Intelligence Group
- AI models
- Gemini
From the original article
3 min read at helpnetsecurity.com
Google’s threat intelligence researchers have linked a zero-day exploit to AI-assisted development by a criminal group. The exploit targeted a popular open-source web-based system administration tool. It allowed attackers to bypass two-factor authentication once they had valid user credentials.
By @helpnetsecurity · helpnetsecurity.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.