CVE-2026-90614
MediumCVSS 6.3Summary
A weakness was identified in FedML-AI FedML up to 0.9.6. It affects the S3Storage.read_model function in fedml/core/distributed/communication/s3/remote_storage.py of the MQTT+S3 Communication Backend component, and manipulation of the s3_key_str argument causes deserialization. Remote exploitation is possible and the project has not responded to the report.
Risk Assessment
A remote attacker can exploit deserialization to execute arbitrary code or compromise the system. The project's lack of response increases the risk of the vulnerability remaining unpatched.
Recommendation
Restrict network access to the MQTT+S3 Communication Backend component and avoid processing untrusted input in s3_key_str. Monitor the project for a fix and apply it urgently.
Other vulnerabilities in FedML
See all- CVE-2026-5536High
FedML-AI FedML up to version 0.8.9 contains a deserialization vulnerability in the sendMessage function of the grpc_server.py file in the gRPC server component. Remote attackers can perform manipulation leading to deserialization, potentially enabling arbitrary code execution.
- CVE-2026-5535Medium
FedML-AI FedML versions up to 0.8.9 contain a path traversal vulnerability in the MQTT Message Handler component in FileUtils.java. An attacker can remotely manipulate the dataSet argument, leading to access to files outside the allowed directory.
Original NVD description (English source)
A weakness has been identified in FedML-AI FedML up to 0.9.6. Affected by this issue is the function S3Storage.read_model of the file fedml/core/distributed/communication/s3/remote_storage.py of the component MQTT+S3 Communication Backend. This manipulation of the argument s3_key_str causes deserialization. Remote exploitation of the attack is possible. The project was informed of the problem early through an issue report but has not responded yet.