Project-MONAI MONAI insecure deserialization in algo_from_pickle

Executive brief

MONAI, a popular framework for medical AI, contains a vulnerability in its 3D segmentation utility. An attacker could provide a specially crafted file that, when processed by the software, allows them to execute arbitrary commands on the underlying system. This could lead to a complete system takeover, data theft, or disruption of medical research operations.

Technical details

A deserialization vulnerability (CWE-502) exists in the `algo_from_pickle` function within `monai/auto3dseg/utils.py`. The function uses the Python `pickle.loads()` method on data read directly from a file without any prior validation or sanitization. An attacker with the ability to provide a malicious `.pkl` file can achieve arbitrary remote code execution (RCE) when the file is processed. The vulnerability is mitigated by the requirement for high privileges and user interaction (loading the file), but it results in a scope change and high impact on confidentiality, integrity, and availability. This issue is resolved in version 1.6.0.

Affected products

• Project-MONAI monai < 1.6.0

Timeline

• 2026-04-03: disclosed: Initial disclosure to Project-MONAI
• 2026-04-07: advisory: GitHub Advisory published
• 2026-04-07: patched: Version 1.6.0 released

References

• https://api.github.com/users/hnking-star
• https://github.com/hnking-star
• https://api.github.com/users/hnking-star/gists%7B/gist_id%7D
• https://api.github.com/users/hnking-star/repos
• https://avatars.githubusercontent.com/u/175230002?v=4
• https://api.github.com/users/hnking-star/events%7B/privacy%7D