Copyright © 2012 Red Hat, Inc. All rights reserved. Important 2026-07-03T20:36:05 keras: Keras: Arbitrary code execution via deserialization vulnerability 8.8 CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H CWE-502

A vulnerability in keras-team/keras version 3.14.0 allows for arbitrary code execution due to improper handling of deserialization in the `Lambda` layer. Specifically, the `_raise_for_lambda_deserialization()` function fails to enforce the safe-mode guard when `safe_mode` is set to `None`, which is the default value when `from_config()` is called outside of a `SafeModeScope` context. This logic error conflates `None` (unset/default-deny) with `False` (explicitly disabled), bypassing the guard and allowing attacker-controlled `marshal` bytecode to be deserialized. Affected call sites include `keras.layers.deserialize(config)`, `keras.models.clone_model(model)`, and any direct invocation of `Lambda.from_config(config)` without an enclosing `SafeModeScope(True)`. This vulnerability can be exploited to achieve arbitrary OS-level code execution in the context of the server or user process.

A flaw was found in the Keras deep learning library. This vulnerability allows a remote attacker to execute arbitrary code on the system by exploiting improper handling of deserialization in the `Lambda` layer. Specifically, a security safeguard designed to prevent unsafe deserialization is bypassed when the `safe_mode` setting is not explicitly enabled, allowing malicious code to be processed. This can lead to complete compromise of the affected server or user process.

This is an Important arbitrary code execution vulnerability in the Keras deep learning library, impacting Red Hat OpenShift AI components. The flaw arises from improper deserialization in the `Lambda` layer, where a security safeguard is bypassed if `safe_mode` is not explicitly enabled. Exploitation requires user interaction, such as loading a specially crafted Keras model. To reduce exposure, only deserialize Keras models from trusted sources. When deserializing models, explicitly enable `safe_mode` by wrapping the deserialization call within a `SafeModeScope(True)` context. This ensures the deserialization safeguard is enforced, preventing the execution of arbitrary code. Red Hat OpenShift AI (RHOAI) Not affected rhoai/odh-kserve-agent-rhel9 Red Hat OpenShift AI (RHOAI) Not affected rhoai/odh-kserve-controller-rhel9 Red Hat OpenShift AI (RHOAI) Not affected rhoai/odh-kserve-router-rhel9 Red Hat OpenShift AI (RHOAI) Affected rhoai/odh-kserve-storage-initializer-rhel9 Red Hat OpenShift AI (RHOAI) Affected rhoai/odh-modelmesh-runtime-adapter-rhel9 Red Hat OpenShift AI (RHOAI) Affected rhoai/odh-pipeline-runtime-tensorflow-cuda-py312-rhel9 Red Hat OpenShift AI (RHOAI) Affected rhoai/odh-pipeline-runtime-tensorflow-rocm-py312-rhel9 Red Hat OpenShift AI (RHOAI) Affected rhoai/odh-workbench-jupyter-tensorflow-cuda-py312-rhel9 Red Hat OpenShift AI (RHOAI) Affected rhoai/odh-workbench-jupyter-tensorflow-rocm-py312-rhel9 https://www.cve.org/CVERecord?id=CVE-2026-12481 https://nvd.nist.gov/vuln/detail/CVE-2026-12481 https://huntr.com/bounties/59ceaed1-c8a3-4135-8f94-169ade02823d