CVE-2024-3660-PoC

CVE-2024-3660 – TensorFlow Keras Arbitrary Code Execution via Malicious Model Overview CVE-2024-3660 is an arbitrary code execution vulnerability in TensorFlow's Keras framework affecting versions \ --build-arg LPORT= \ -t tfimg . && \ containerid=$docker create tfimg && \ docker cp...

TensorFlow CVE-2024-3660 Keras H5 Lambda RCE

CVE-2024-3660: crafted Keras .h5 with Lambda layer deserializes marshal code via legacy loadmodel path when ML service loads uploads. !/usr/bin/env python3 -- coding: utf-8 -- """TensorFlow/Keras CVE-2024-3660 malicious H5 Lambda layer RCE.""" from future import annotations import codecs import...

PT-2026-76464

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...

Exploit for Execution with Unnecessary Privileges in Oracle Jdk

--- CVE-2026-22008 – AWS Lambda Layer Injection via Shared L...

CVE-2026-12481

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 bypasse...

Linux Distros Unpatched Vulnerability : CVE-2026-12481

The Linux/Unix host has one or more packages installed that are impacted by a vulnerability without a vendor supplied patch available. - 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...

DEBIAN-CVE-2026-12481

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 raiseforlambdadeserialization function fails to enforce the safe-mode guard when safemode is set to None, which is the default...

CVE-2026-12481

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 raiseforlambdadeserialization function fails to enforce the safe-mode guard when safemode is set to None, which is the default...

UBUNTU-CVE-2026-12481

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 raiseforlambdadeserialization function fails to enforce the safe-mode guard when safemode is set to None, which is the default...

CVE-2026-12481 Deserialization of Untrusted Data in keras-team/keras

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 raiseforlambdadeserialization function fails to enforce the safe-mode guard when safemode is set to None, which is the default...

CVE-2026-12481 Deserialization of Untrusted Data in keras-team/keras

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 raiseforlambdadeserialization function fails to enforce the safe-mode guard when safemode is set to None, which is the default...

CVE-2026-12481

The CVE affects keras-team/keras 3.14.0, where deserialization in the Lambda layer can bypass the safe-mode guard. The root cause is _raise_for_lambda_deserialization() treating safe_mode=None as unset/deny, allowing attacker-controlled marshal bytecode to be deserialized. Affected call sites inc...

CVE-2026-12481

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 raiseforlambdadeserialization function fails to enforce the safe-mode guard when safemode is set to None, which is the default...

CVE-2026-12481 Deserialization of Untrusted Data in keras-team/keras

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 raiseforlambdadeserialization function fails to enforce the safe-mode guard when safemode is set to None, which is the default...

CVE-2026-12481

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 "raiseforlambdadeserialization" function fails to enforce the safe-mode guard when "safemode" is set to "None", which is the...

CVE-2026-12481

Removed by vendor...

PT-2026-55585

Name of the Vulnerable Software and Affected Versions keras-team/keras version 3.14.0 Description Improper handling of deserialization in the Lambda layer allows for arbitrary OS-level code execution in the context of the server or user process. The raise for lambda deserialization function fails...

CVE-2026-12481: Deserialization of Untrusted Data

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 raiseforlambdadeserialization function fails to enforce the safe-mode guard when safemode is set to None, which is the default...

Lambda Layer `safe_mode=None` Guard Bypass — Arbitrary Code Execution via `from_config()`

Description Keras's Lambda layer deserializes Python functions using marshal.loads on attacker-controlled bytecode. A safe-mode guard in raiseforlambdadeserialization is intended to block this path, but a logic error causes the guard to silently pass when safemode is None — the default value when...

Arbitrary Code Execution

Keras is vulnerable to Arbitrary Code Execution. The vulnerability is due to Model.loadmodel not honoring safemode=True when reading legacy .h5/.hdf5 archives and deserializing pickled Lambda-layer code from a crafted model file, which allows an attacker to supply a malicious archive that execute...