Security
Amazon says a suspected China-linked operator used autonomous AI agents to find and exploit a MONAI Label zero-day and gain root on a university server
Amazon says Cairn AI agents, run by a suspected China-linked operator, found a MONAI Label zero-day in about 11 minutes.
Amazon Threat Intelligence said on 20 November that autonomous AI agents running on the open-source Cairn framework found and exploited a previously unknown flaw in MONAI Label, a medical-imaging server, and gained root access to a university server in East Asia. The attribution to a China-linked operator is made with moderate confidence.
The essentials
- What happened: over about 48 hours from 8 September, the agents mapped the application and reached file writing in roughly 11 minutes. In a separate session they planted a Python file that ran at startup and got a root shell, per The Hacker News.
- Who set the goal: the operator asked for shell access and chose the targets; Amazon says the agents picked the methods. Evidence for the China link is Simplified Chinese prompts, China Standard Time settings and the targets, and no named group or government role was established.
- The flaw: a path-traversal bug (CVE-2026-133968, rated 9.8) in MONAI Label 0.8.5 and earlier, where authentication is off by default. Maintainers merged a source fix on 19 October, but the advisory lists no patched release yet.
- Scope: one university server is confirmed compromised. A second university network was reportedly accessed through a tunnel the operator provided, which Amazon believes but has not confirmed.
How Amazon reconstructed it
The operator’s Cairn control panel was exposed online without a password, which revealed task settings, prompts and run history. The account rests on Amazon’s own reconstruction and has not been independently confirmed.
What is not established
- Whether patient scans or records were accessed or taken: not stated in the sources.
- A named threat group or government sponsorship: not stated in the sources.
- Total number of universities or servers affected: not stated in the sources.
Summary table
| Point | What is stated as of November 21, 2026 | Source |
|---|---|---|
| Campaign window | 8 to 10 September 2026 | Amazon / The Hacker News |
| Affected versions | MONAI Label 0.8.5 and earlier | The Hacker News |
| Attribution confidence | Moderate (China-nexus) | The Hacker News |
FAQ
Is there a patch for the MONAI Label flaw?
A source fix was merged on 19 October, but the advisory lists no patched release and the latest PyPI version is still 0.8.5.
Sources
- The Hacker News: Attackers’ AI Agents Found a Zero-Day in Medical-Imaging Software and Gained Root Access (verified 2026-11-21)
- Cryptopolitan: Amazon urges automated defenses after AI agents hacked a university (verified 2026-11-21)
- World Today Journal: Amazon: AI Agents Exploited MONAI Label Zero-Day in 11 Minutes; No Patch Released (verified 2026-11-21)
- WindowsForum: Amazon: Cairn AI Agents Exploited MONAI Label Zero-Day to Gain Root Access (verified 2026-11-21)
Spotted an error in this article? Report it · Our corrections policy