Логотип exploitDog
Консоль
Логотип exploitDog

exploitDog

nvd логотип

CVE-2026-1462

Опубликовано: 13 апр. 2026
Источник: nvd
CVSS3: 8.8
CVSS3: 7.8
EPSS Низкий

Описание

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 safe_mode=True. This bypasses the security guarantees of safe_mode and enables arbitrary attacker-controlled code execution during model inference under the victim's privileges. The issue arises due to the unconditional loading of external SavedModels, serialization of attacker-controlled file paths, and the lack of validation in the from_config() method.

Уязвимые конфигурации

Конфигурация 1
cpe:2.3:a:keras:keras:3.13.0:*:*:*:*:*:*:*
Конфигурация 2
cpe:2.3:a:redhat:openshift_ai:*:*:*:*:*:*:*:*
Версия от 2.25 (включая) до 2.25.7 (исключая)

EPSS

Процентиль: 34%
0.00405
Низкий

8.8 High

CVSS3

7.8 High

CVSS3

Дефекты

CWE-502
CWE-502

Связанные уязвимости

CVSS3: 7.8
ubuntu
4 месяца назад

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 `safe_mode=True`. This bypasses the security guarantees of `safe_mode` and enables arbitrary attacker-controlled code execution during model inference under the victim's privileges. The issue arises due to the unconditional loading of external SavedModels, serialization of attacker-controlled file paths, and the lack of validation in the `from_config()` method.

CVSS3: 7.8
redhat
4 месяца назад

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 `safe_mode=True`. This bypasses the security guarantees of `safe_mode` and enables arbitrary attacker-controlled code execution during model inference under the victim's privileges. The issue arises due to the unconditional loading of external SavedModels, serialization of attacker-controlled file paths, and the lack of validation in the `from_config()` method.

CVSS3: 7.8
debian
4 месяца назад

A vulnerability in the `TFSMLayer` class of the `keras` package, versi ...

CVSS3: 8.8
github
4 месяца назад

Keras has an untrusted deserialization vulnerability

EPSS

Процентиль: 34%
0.00405
Низкий

8.8 High

CVSS3

7.8 High

CVSS3

Дефекты

CWE-502
CWE-502