AI agents used a MONAI Label zero-day to gain root on a university server
Original title: Attackers' AI Agents Found a Zero-Day in Medical-Imaging Software and Gained Root Access
Amazon says unauthenticated path traversal in 0.8.5 and earlier planted a Python startup file; no installable fix exists.
Amazon Threat Intelligence said suspected China-linked operators used autonomous AI agents to find a path-traversal flaw in MONAI Label and gain root on one university server. CVE-2026-133968 lets an unauthenticated client write files outside the storage folder on versions 0.8.5 and earlier, which default to no login. Maintainers merged a source fix on October 19 but have not published an installable release. Amazon has not confirmed a second university compromise or theft of scans.
Unconfirmed: Amazon's reconstruction is unconfirmed independently; the second university compromise and any government role are not established.
- Amazon found an exposed, passwordless Cairn panel with agent prompts and history.
- Agents reached file writes in about 11 minutes during a September session.
- Root came later when a planted Python startup file loaded during training.
- The CVE is rated 9.8; the project advisory still lists no patched version.
- China attribution is moderate confidence and is not tied to a named group.
Impact: Unauthenticated file write outside storage, leading to root code execution
| Product | Affected |
|---|---|
| MONAI Label | 0.8.5 and earlier |
- 1.Block untrusted network access to MONAI Label with firewall or routing rules.
- 2.Enable authentication, which is off by default, and run the server as a limited account.
- 3.If exposed, check for unexpected Python startup files and unusual training jobs.
- CVEs
- CVE-2026-133968
- Products
- MONAI Label
- Malware
- Cairn
- Organizations
- Amazon Threat Intelligence
- Countries
- China
Vulnerabilities mentioned
All| CVE | Vulnerability | CVSS | EPSS | Signals | Published |
|---|---|---|---|---|---|
| CVE-2026-133968 | Directory traversal to RCE in MONAI Label through 0.8.5 | 9.8 | — |
From the original article
4 min read at thehackernews.com
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.
thehackernews.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.