CVE-2026-108958: TensorLayer HDF5 Loading utils.py load_hdf5_graph deserialization
A weakness has been identified in TensorLayer up to 2.2.4. This affects the function loadhdf5graph of the file tensorlayer/files/utils.py of the component HDF5 Loading. Executing a manipulation of the argument modelconfig can lead to deserialization. The attack can be executed remotely. The exploit has been made available to the public and could be used for attacks. The project was informed of the problem early through an issue report but has not responded yet.
Affected Software
Event History
Frequently Asked Questions
Who is exposed to this issue?
Deployments using TensorLayer versions up to 2.2.4 that load HDF5 graph data through the affected load_hdf5_graph function are exposed. The issue can be exploited remotely, but requires a user to interact with manipulated model_config input.
What does an attacker need to exploit it?
An attacker needs to provide or cause processing of a manipulated model_config argument during HDF5 graph loading. No privileges are required, and public exploit code is available.
Is a fix available from the project?
The available information does not identify a fixed version. The project was reportedly notified through an issue report but had not responded at the time of publication.
What can be done if upgrading is not immediately possible?
Avoid loading HDF5 models or graph data from untrusted sources through load_hdf5_graph, and restrict who can supply or modify model_config values. Because exploitation requires user interaction, prevent users or automated workflows from accepting untrusted model artifacts.