China-linked AI agents exploited a MONAI Label zero-day, Amazon says
Original title: Amazon: China-Linked Hackers’ AI Agents Exploited Medical Software Zero-Day
Agents wrote a Python startup file on university servers and obtained a root shell; versions 0.8.5 and earlier are affected.
Amazon says autonomous AI agents operated by a China-linked actor found and exploited a path-traversal zero-day in MONAI Label, an open-source tool for annotating medical images. Amazon detected the activity in September through an unauthenticated Cairn instance whose logs showed a roughly 48-hour campaign against university addresses in East Asia outside mainland China. Developers rate the bug critical in versions 0.8.5 and earlier, with authentication off by default, and say a fix will ship in the next release.
Unconfirmed: Amazon infers the second university was compromised with the same zero-day, and attributes the actor to a Chinese nexus only with moderate confidence.
- Agents went from mapping an API to writing files in about 11 minutes, Amazon said.
- A later session planted a Python reverse shell and achieved a root shell.
- A proxy then tunneled into a second university running the same software.
- Amazon assesses the flaw was a zero-day with high confidence.
- Attribution to a Chinese-nexus actor is moderate confidence only.
Impact: Arbitrary file write leading to a root shell
| Product | Affected |
|---|---|
| MONAI Label | 0.8.5 and earlier |
- 1.Do not expose MONAI Label 0.8.5 and earlier with default authentication disabled.
- 2.Install the fix when it ships in the next MONAI Label release.
- Vendors
- MONAI
- Products
- MONAI LabelCairn
- Organizations
- Amazon
- Countries
- China
From the original article
2 min read at securityweek.com
A threat actor used autonomous Cairn AI agents to discover and exploit a critical MONAI Label vulnerability. The post Amazon: China-Linked Hackers’ AI Agents Exploited Medical Software Zero-Day appeared first on SecurityWeek.
By Eduard Kovacs · securityweek.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.