Логотип exploitDog
Консоль
Логотип exploitDog

exploitDog

nvd логотип

CVE-2026-5817

Опубликовано: 22 мая 2026
Источник: nvd
CVSS3: 8.2
CVSS3: 8.6
EPSS Низкий

Описание

The vllm-metal inference backend in Docker Model Runner on macOS unconditionally sets trust_remote_code=True when loading model tokenizers, and runs without sandboxing. This causes transformers.AutoTokenizer.from_pretrained() to import and execute arbitrary Python files included in any model pulled from an OCI registry, resulting in arbitrary code execution on the Docker host as the Docker Desktop user when inference is triggered.

Any container on the Docker network can trigger this by calling the model-runner.docker.internal API to pull a malicious model and request inference.

Уязвимые конфигурации

Конфигурация 1

Одновременно

cpe:2.3:a:docker:docker_desktop:*:*:*:*:*:*:*:*
Версия от 4.62.0 (включая) до 4.68.0 (исключая)
cpe:2.3:o:apple:macos:-:*:*:*:*:*:*:*

EPSS

Процентиль: 13%
0.00224
Низкий

8.2 High

CVSS3

8.6 High

CVSS3

Дефекты

CWE-829

Связанные уязвимости

CVSS3: 8.2
github
3 месяца назад

The vllm-metal inference backend in Docker Model Runner on macOS unconditionally sets trust_remote_code=True when loading model tokenizers, and runs without sandboxing. This causes transformers.AutoTokenizer.from_pretrained() to import and execute arbitrary Python files included in any model pulled from an OCI registry, resulting in arbitrary code execution on the Docker host as the Docker Desktop user when inference is triggered. Any container on the Docker network can trigger this by calling the model-runner.docker.internal API to pull a malicious model and request inference.

CVSS3: 8.6
fstec
3 месяца назад

Уязвимость компонента vllm-metal inference backend плагина для Docker Desktop Docker Model Runner, позволяющая нарушителю выполнить произвольный код

EPSS

Процентиль: 13%
0.00224
Низкий

8.2 High

CVSS3

8.6 High

CVSS3

Дефекты

CWE-829