Attackers' AI Agents Found a Zero-Day in Medical-Imaging Software and Gained Root Access
AI Summary
Suspected China-linked attackers used autonomous AI agents to discover and exploit a zero-day path traversal vulnerability in MONAI Label, an open-source medical imaging server, gaining root access on a university server. The flaw, tracked as CVE-2026-133968, allows unauthenticated attackers to write files outside the storage directory due to insufficient path validation, enabling privilege escalation via malicious Python startup files. Although a fix has been merged into the source code, no official patched version has been released, leaving systems vulnerable. The campaign leveraged the open-source AI penetration testing framework Cairn, with evidence including Simplified Chinese prompts and China Standard Time settings pointing to a China-linked actor, though no specific group was named.
AI-extracted · verify before operational use