CVE-2026-5817

Published May 22, 2026

Last updated 6 days ago

Overview

Description
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.
Source
security@docker.com
NVD status
Analyzed
Products
docker_desktop

Risk scores

CVSS 4.0

Type
Secondary
Base score
8.8
Impact score
-
Exploitability score
-
Vector string
CVSS:4.0/AV:L/AC:L/AT:P/PR:L/UI:N/VC:H/VI:H/VA:H/SC:H/SI:H/SA:H/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X
Severity
HIGH

CVSS 3.1

Type
Primary
Base score
8.6
Impact score
6
Exploitability score
1.8
Vector string
CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:C/C:H/I:H/A:H
Severity
HIGH

Weaknesses

security@docker.com
CWE-829

Social media

Hype score
Not currently trending

Configurations

References

Sources include official advisories and independent security research.