CVE-2026-1462

A vulnerability in the TFSMLayer class of the keras package, version 3.13.0, allows attacker-controlled TensorFlow SavedModels to be loaded during deserialization of .keras models, even when safemode=True. This bypasses the security guarantees of safemode and enables arbitrary attacker-controlled...

UBUNTU-CVE-2026-1462

A vulnerability in the TFSMLayer class of the keras package, version 3.13.0, allows attacker-controlled TensorFlow SavedModels to be loaded during deserialization of .keras models, even when safemode=True. This bypasses the security guarantees of safemode and enables arbitrary attacker-controlled...

CVE-2026-1462

A vulnerability in the TFSMLayer class of the keras package, version 3.13.0, allows attacker-controlled TensorFlow SavedModels to be loaded during deserialization of .keras models, even when safemode=True. This bypasses the security guarantees of safemode and enables arbitrary attacker-controlled...

CVE-2026-1462 Safe Mode Bypass in keras-team/keras

A vulnerability in the TFSMLayer class of the keras package, version 3.13.0, allows attacker-controlled TensorFlow SavedModels to be loaded during deserialization of .keras models, even when safemode=True. This bypasses the security guarantees of safemode and enables arbitrary attacker-controlled...

CVE-2026-1462

The CVE-2026-1462 issue affects the keras package (v3.13.0) via the TFSMLayer deserialization path. The vulnerability allows attacker-controlled TensorFlow SavedModels to be loaded during deserialization of .keras models, even when safe_mode is enabled, due to unconditional loading of external Sa...

CVE-2026-1462

Removed by vendor...

CVE-2026-1462 Safe Mode Bypass in keras-team/keras

A vulnerability in the TFSMLayer class of the keras package, version 3.13.0, allows attacker-controlled TensorFlow SavedModels to be loaded during deserialization of .keras models, even when safemode=True. This bypasses the security guarantees of safemode and enables arbitrary attacker-controlled...

CVE-2026-1462

A vulnerability in the TFSMLayer class of the keras package, version 3.13.0, allows attacker-controlled TensorFlow SavedModels to be loaded during deserialization of .keras models, even when safemode=True. This bypasses the security guarantees of safemode and enables arbitrary attacker-controlled...

PT-2026-32367

Name of the Vulnerable Software and Affected Versions keras version 3.13.0 Description A flaw in the TFSMLayer class allows attacker-controlled TensorFlow SavedModels to be loaded during the deserialization of .keras models. This occurs even when safe mode=True is enabled, bypassing security...

Keras 代码问题漏洞

Keras is an open-source deep learning framework with multiple backends. Version 3.13.0 of Keras contains a code vulnerability that stems from the TFSLayer class’s unconditional loading of external SavedModels, which may lead to arbitrary code execution...

Security Bulletin: Arbitrary File Read, SSRF, and Code Execution Vulnerabilities in TensorFlow Keras Model Loading (v2.13) affects watsonx.data

Summary A vulnerability in TensorFlow Keras v2.13 allows malicious .keras model files to trigger arbitrary local file reads, Server-Side Request Forgery SSRF, and potential code execution during model loading—even when safemode=True is enabled. The issue arises from improper handling of external...

[CWE-73] Keras — Arbitrary HDF5 File Read via Virtual Dataset Bypass of External Storage Check (CVE-2026-1669 Incomplete Fix)

Description The fix for CVE-2026-1669 in Keras's H5IOStore.verifydataset method savinglib.py:1051 is incomplete. The patch checks dataset.external to block HDF5 datasets that use external file storage, but it does not check dataset.isvirtual. HDF5 Virtual Datasets VDS — created via...

📄 Google Keras 3.13.0 Denial of Service

A denial of service vulnerability exists in the HDF5 weight loading component of Google Keras versions 3.0.0 through 3.13.0 on all platforms. The vulnerability is caused by the absence of any validation or throttling when processing HDF5 dataset shape metadata declared inside a .keras archive...

Path Traversal in DiskIOStore via Unsanitized Layer Names in Keras 3

Description A logic flaw in the Keras 3 v3.0.0+ model saving and loading library handles internal asset paths insecurely. Specifically, the DiskIOStore.make method in keras/src/saving/savinglib.py uses user-provided layer names to construct directory paths without sanitizing for parent directory...

Directory Traversal

Keras is vulnerable to Directory Traversal. The vulnerability is due to unsafe extraction of tar archives in keras.utils.getfile without proper filtering during extraction, which allows an attacker to bypass path validation and write files outside the intended directory...

Path Traversal in Keras Archive Extraction via CWD Validation Bypass Leading to Arbitrary File Write

Description Technical Details of the Vulnerability Summary Keras's archive extraction utilities in keras/src/utils/fileutils.py are vulnerable to path traversal. The functions filtersafetarinfos and filtersafezipinfos validate archive member paths against the process current working directory...

model.weights.h5: h5py.ExternalLink at Group level silently followed during load_model(), bypassing CVE-2025-9905 fix — information disclosure from arbitrary HDF5 files

Keras 3.x introduced a fix for CVE-2025-9905 by checking dataset.external in H5IOStore.verifydataset. This check blocks datasets whose raw bytes are stored in external files via the HDF5 "External Data Storage" mechanism. However, HDF5 supports a second, unrelated external-reference mechanism:...

Important: Red Hat Security Advisory: RHTAS 1.3.2 - Tech Preview Release of Model Transparency

The Tech Preview release of the RHTAS Model Transparency CLI image. For more details please visit the product documentation at https://access.redhat.com/documentation/en-us/redhattrustedartifactsigner/1.3 The RHTAS Model Transparency CLI image can be used to sign and verify AI/ML workloads...

AZL-79362 CVE-2026-3381 affecting package keras 2.11.0-3

Compress::Raw::Zlib versions through 2.219 for Perl use potentially insecure versions of zlib. Compress::Raw::Zlib includes a copy of the zlib library. Compress::Raw::Zlib version 2.220 includes zlib 1.3.2, which addresses findings fron the 7ASecurity audit of zlib. The includes fixs for...

Security Bulletin: IBM Maximo Application Suite - Maximo AI Service uses multiple third party dependencies which are vulnerable to multiple CVEs.

Summary IBM Maximo Application Suite - Maximo AI Service uses "torch-2.9.1-cp311-cp311-manylinux228x8664.whl, keras-3.12.0-py3-none-any.whl, hibernate-core-6.6.36.Final.jar" dependencies which are vulnerable to "CVE-2025-2998, CVE-2025-2999, CVE-2025-55552, CVE-2025-63396, CVE-2026-0897,...