This vulnerability affects the FedML-AI FedML project up to version 0.9.6 in the function S3Storage.read_model located in fedml/core/distributed/communication/s3/remote_storage.py. The issue arises from manipulation of the s3_key_str argument, which leads to deserialization. Remote exploitation is possible without user interaction and requires low attack complexity with low privileges. The vulnerability has a CVSS 4.0 base score of 5.3 (medium severity). The vendor has been notified but has not yet responded. Patch availability is indicated, and as a cloud service, remediation is typically managed by the vendor.

The vulnerability allows remote attackers to exploit deserialization in the S3Storage.read_model function, potentially leading to unauthorized code execution or other impacts associated with unsafe deserialization. The attack requires low privileges and no user interaction. The overall impact is rated medium severity with limited confidentiality, integrity, and availability impact.

A patch is available for this vulnerability. Since FedML is a cloud-hosted service, the vendor typically manages remediation server-side. Users should monitor the vendor advisory for confirmation of patch deployment and apply any recommended updates or mitigations once available.