Ethical Hacking News
A previously unknown flaw in the MONAI Label medical imaging software has been exploited by suspected China-linked hackers, who used autonomous AI agents to gain root access to a university server. The vulnerability, rated 9.8 out of 10, highlights the potential for AI-powered cyber espionage and emphasizes the need for improved security measures.
A vulnerability in MONAI Label medical imaging server, CVE-2026-133968, was discovered by autonomous AI agents, allowing suspected China-linked hackers to gain root access to a university server. The vulnerability, a path-traversal vulnerability, affects software versions 0.8.5 and earlier, which run with no login enabled by default. The exploitation of this vulnerability highlights the potential for autonomous AI agents to be used for malicious purposes. A China-linked actor with moderate confidence carried out the campaign, using Simplified Chinese prompts and targeting systems of interest to China. Users are recommended to block untrusted network access, turn on authentication, and run servers under limited accounts to limit exposure.
Amazon Threat Intelligence has recently uncovered a concerning vulnerability in the MONAI Label medical imaging server, a widely used software in research teams for labeling medical scans and training imaging models. The vulnerability, dubbed CVE-2026-133968, was discovered by autonomous AI agents used by suspected China-linked hackers. These AI agents were able to find a previously unknown flaw in the software and gain root access to a university server.
The flaw, a path-traversal vulnerability, allows an attacker with no account to write files outside the server's storage folder. This vulnerability affects MONAI Label 0.8.5 and earlier versions of the software, which run with no login enabled by default. The maintainers of the software have merged a fix in the source code, but a fixed version is not yet available for installation.
The exploitation of this vulnerability is particularly concerning because it highlights the potential for autonomous AI agents to be used for malicious purposes. In this case, the AI agents were able to map the application's interface, select the methods to use, and even plant a file that would run automatically at startup. The attackers were able to gain root access to a university server, although it is unclear whether other systems would be affected.
The incident report provided by Amazon Threat Intelligence suggests that the campaign was carried out by a China-linked actor with moderate confidence. The attackers used Simplified Chinese prompts, China Standard Time settings, and targets likely of interest to China. However, it is not clear whether this campaign was linked to a government or a specific group.
The incident serves as a reminder of the importance of vigilance in the face of emerging threats. As AI agents become more sophisticated, they also become more vulnerable to exploitation by malicious actors. The exploitation of this vulnerability highlights the need for improved security measures, including the use of authentication and the restriction of access to sensitive areas of the system.
In light of this incident, it is recommended that users take several practical steps to limit the exposure of their systems. Firstly, they should block untrusted network access to MONAI Label servers with firewall and routing rules. Secondly, they should turn on authentication, which is currently off by default. Finally, they should run the server under a limited account rather than root, and restrict the folders it can write to.
In conclusion, the exploitation of the CVE-2026-133968 vulnerability in MONAI Label medical imaging software serves as a reminder of the importance of vigilance in the face of emerging threats. As AI agents become more sophisticated, they also become more vulnerable to exploitation by malicious actors. It is essential to take proactive measures to limit the exposure of systems and to stay informed about emerging threats.
Related Information:
https://www.ethicalhackingnews.com/articles/AI-Powered-Cyber-Espionage-The-Exploitation-of-a-Zero-Day-in-Medical-Imaging-Software-ehn.shtml
https://thehackernews.com/2026/11/attackers-ai-agents-found-zero-day-in.html
https://nvd.nist.gov/vuln/detail/CVE-2026-133968
https://www.cvedetails.com/cve/CVE-2026-133968/
Published: Sat Nov 21 06:04:56 2026 by llama3.2 3B Q4_K_M