Cyber Resilience

CVE-2026-5817

HighUpdated

Published: 22 May 2026

Published
22 May 2026
Modified
01 June 2026
KEV Added
Patch
CVSS Score v4 8.8 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
EPSS Score 0.0022 13.1th percentile
Risk Priority 55 floored blend · peak EPSS

Summary

CVE-2026-5817 is a high-severity Inclusion of Functionality from Untrusted Control Sphere (CWE-829) vulnerability in Docker Docker Desktop. Its CVSS base score is 8.8 (High).

Operationally, exploitation aligns with the MITRE ATT&CK technique Exploit Public-Facing Application (T1190); ranked at the 13.1th percentile by exploit likelihood (below the median); it is not currently listed in the CISA KEV catalog.

This vulnerability is AI-related — categorised as NLP and Transformers; in the Supply Chain and Deployment risk domain.

OWASP Top 10 for Web (2025)

EU & UK References

Vulnerability details

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,…

more

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.

CWE(s)

AI Security AnalysisAI

AI Category
NLP and Transformers
Risk Domain
Supply Chain and Deployment
OWASP Top 10 for LLMs 2025
None mapped
Classification Reason
Matched keywords: transformers, vllm

Related Threats

MITRE ATT&CK Enterprise TechniquesAI

T1190 Exploit Public-Facing Application Initial Access
Adversaries may attempt to exploit a weakness in an Internet-facing host or system to initially access a network.
T1059.006 Python Execution
Adversaries may abuse Python commands and scripts for execution.
Why these techniques?

Direct RCE via exposed model-runner API (T1190) executing attacker-controlled Python from untrusted OCI models (T1059.006).

Confidence: HIGH · MITRE ATT&CK Enterprise v19.0

Affected Assets

docker
docker desktop
4.62.0 — 4.68.0

Mitigating Controls

Likely Mitigating Controls AI

Per-CVE control mapping for this CVE has not run yet; the list below is derived from the weakness types (CWEs) cited in the NVD entry.

addresses: CWE-829

Limiting P2P file sharing technology reduces inclusion of functionality or resources from untrusted external control spheres.

addresses: CWE-829

Enforcing installation policies prevents users from including functionality obtained from untrusted control spheres.

addresses: CWE-829

The inventory process requires identifying and recording the origin of all components, making inclusion of functionality from untrusted control spheres easier to detect during reviews.

addresses: CWE-829

Requiring approval and monitoring of maintenance tools prevents inclusion and execution of functionality obtained from untrusted sources.

addresses: CWE-829

Unowned portable devices represent untrusted control spheres; the prohibition prevents inclusion of functionality or data from such sources.

addresses: CWE-829

Strategy mandates assessment of third-party components and suppliers, directly reducing inclusion of functionality from untrusted control spheres.

addresses: CWE-829

Procedures can mandate supply-chain vetting and restrictions on functionality obtained from untrusted third-party or external control spheres.

addresses: CWE-829

Requires use of trusted sources and provenance tracking, tangibly limiting inclusion of functionality from untrusted control spheres.

References