{
  "id": 21042729,
  "title": "Attackers' AI Agents Found a Zero-Day in Medical-Imaging Software and Gained Root Access",
  "url": "https://urgent.news/2026/11/21/attackers-ai-agents-found-a-zero-day-in-medical-imaging-software-and",
  "topic": "ai",
  "section": "AI",
  "published": "2026-11-21T11:03:17.000Z",
  "source": {
    "name": "The Hacker News",
    "slug": "the-hacker-news",
    "url": "https://thehackernews.com/2026/11/attackers-ai-agents-found-zero-day-in.html"
  },
  "original_language": "en",
  "account": "Suspected China-linked hackers employed autonomous AI agents to uncover a previously unknown vulnerability in MONAI Label, an open-source medical imaging server, and gain root access to a university server, according to Amazon Threat Intelligence. The flaw enables an unauthorized user to write files outside the server's storage folder, affecting MONAI Label versions 0.8.5 and earlier in the default setup, which runs without login enabled. MONAI Label is utilized by research teams to label medical scans and train imaging models; it operates separately from the broader MONAI framework. The weakness stems from a path-traversal flaw in the server's storage of uploaded images and labels, where a file identifier is appended to the storage folder and written to disk without verifying that the result remains within the designated folder. The server operates with the privileges of the account it runs under, so the attacker cannot access the entire machine, but a writable code or settings location can lead to executing code. Authentication is off by default, allowing any client to set off the write. Amazon traced the intrusion by discovering the operator's Cairn control panel on the internet without a password. The AI agents mapped the application's interface and wrote files to the server within roughly 11 minutes, working with minimal guidance. The operator specified the goal of obtaining shell access and selected the targets, while the agents determined the methods. Following this, the agents planted a file that Python runs automatically at startup, triggering a training feature that started a new Python process, which loaded the malicious file and established a connection to the attacker as root. This root access was achieved on one university server, and whether other MONAI Label systems would yield the same result depends on their configuration and service account permissions. The maintainers confirmed the flaw and merged the source fix on October 19, which prevents uploads from resolving outside the approved folders and rejects empty names or names starting with a dot. However, there is no patched version available, as the correction is expected in a future release. The vulnerability has been cataloged as CVE-2026-133968, rated as critical (score 9.8 out of 10). The practical steps to mitigate the issue include cutting off exposure and limiting the potential reach of a compromised server, as patient scans were not accessed, clinical systems were not touched, and the attacker did not extend beyond the university targets described in the incident.",
  "summary": "Suspected China-linked hackers used autonomous AI agents to find a previously unknown flaw in MONAI Label, an open-source medical imaging server, and gain root access to a university server, Amazon Threat Intelligence said this week. The flaw allows an attacker with no account to write files outside the server's storage folder, and it affects MONAI Label 0.8.5 and earlier in the default setup,",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}