A high-severity path traversal vulnerability in the MONAI AI toolkit, tracked as CVE-2025-58755, allows remote attackers to overwrite arbitrary system files.

The application utilizes the zip_file.extractall function without proper validation, enabling path traversal. This flaw allows an authenticated user to craft a malicious ZIP archive that, when processed, can write files to unintended directories on the host system.

The vulnerability carries a CVSS score of 8.8, reflecting its potential for total system compromise. Successful exploitation could lead to arbitrary code execution or the corruption of critical system files, resulting in significant service disruption, loss of data integrity, and potential exposure of sensitive medical imaging information.

Immediate Action: Update the MONAI package to version 1.5.1 or later immediately to incorporate the necessary input validation fixes.

Proactive Monitoring: Monitor system logs for unusual file write activity or access attempts directed toward directories outside of the designated working path for the MONAI toolkit.

Compensating Controls: Implement strict file system permissions for the service account running MONAI to limit the scope of potential file overwrites, and utilize security scanning tools to inspect incoming ZIP archives for path traversal patterns before processing.

Public Exploit Available: Unknown.

Given the severity of this path traversal vulnerability and the potential for full system compromise, organizations should prioritize updating their MONAI instances to version 1.5.1. Administrators must ensure that all environments utilizing this toolkit are patched, as the lack of path sanitization during decompression presents a critical risk to the underlying host infrastructure.