Multiple MONAI Vulnerabilities Lead to Remote Code Execution Risks

• •MONAI versions prior to 1.6.0 are vulnerable to multiple RCE flaws.
• •Exploitation methods include OS command injection and attribute chain bypass.
• •Immediate updates to MONAI are recommended to mitigate risks.

A series of vulnerabilities affecting MONAI versions prior to 1.6.0 have been disclosed, allowing for remote code execution (RCE) through various attack vectors including OS command injection and pickle deserialization. The vulnerabilities include CVE-2026-1234, CVE-2026-1235, and CVE-2026-1236, impacting users who utilize MONAI for medical imaging and AI applications. Attackers can exploit these vulnerabilities via specially crafted dataset names, attribute chains, or configuration files. The affected versions are MONAI 1.5.2 and earlier, with the latest version 1.6.0 addressing some of these issues. Security advisories have been issued, urging immediate updates to mitigate risks. The scope of impact is significant, given MONAI's use in healthcare and research environments. Current status indicates that exploitation is possible, with proof-of-concept code available for some vulnerabilities.