Malware Installation via AI Coding Assistants Exploited

• •Hackers are using AI coding assistants to install malware via malicious text files.
• •Files can be dropped into Downloads folders without user consent, allowing exploitation.
• •The attack can lead to persistent malware installation across projects on affected systems.

Security firm Forever reported that hackers are exploiting AI coding assistants by using malicious instructions hidden in text files located in users' Downloads folders. These files can instruct the assistants to install malware, specifically linked to the MALFEX campaign. The attack method involves dropping files into the Downloads folder without user consent, allowing the AI to execute harmful commands. Forever demonstrated that when the AI assistant, such as OpenAI's Codex, processes these files, it can inadvertently install malware and alter system settings to maintain persistent access. The researchers noted that this method could compromise any project on the affected machine. While the exact method of file delivery remains unclear, the incident highlights the risks of AI assistants operating with extensive permissions. Forever's findings follow similar research demonstrating the potential for AI tools to be manipulated for malicious purposes.