CVE-2026-5817
HIGHDescription
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.
CVSS v3.1 Score
EPSS — Exploit Prediction
EPSS estimates the probability that this vulnerability will be exploited in the wild within the next 30 days. A higher score means more likely to be exploited.
Weakness Type (CWE)
References
Other References
Frequently Asked Questions
What is CVE-2026-5817? +
How severe is CVE-2026-5817? +
How do I check if I'm vulnerable to CVE-2026-5817? +
Related Vulnerabilities
Cursor is a code editor built for programming with AI. In versions prior to 3.0.0, the Cursor Desktop could execute …
conda-forge-metadata provides programatic access to conda-forge's metadata. conda-forge-metadata uses an optional dependency - "conda-oci-mirror" which was neither present on the …
Collabora Online is a collaborative online office suite based on LibreOffice. Macro support is disabled by default in Collabora Online, …
An Inclusion of Functionality from Untrusted Control Sphere vulnerability in the SSH server on B&R APROL <4.4-00P1 may allow an …
A critical security vulnerability exists in remote cache extensions for common build systems utilizing bucket-based remote cache (such as those …
A PHP objection injection vulnerability exists in the Monero Project’s Laravel-based forum software due to unsafe handling of untrusted input …