CVE-2026-27893
EUVD-2026-1647827.03.2026, 00:16
vLLM is an inference and serving engine for large language models (LLMs). Starting in version 0.10.1 and prior to version 0.18.0, two model implementation files hardcode `trust_remote_code=True` when loading sub-components, bypassing the user's explicit `--trust-remote-code=False` security opt-out. This enables remote code execution via malicious model repositories even when the user has explicitly disabled remote code trust. Version 0.18.0 patches the issue.Enginsight
Affected Products (NVD)
| Vendor | Product | Version |
|---|---|---|
| vllm | vllm | 0.10.1 ≤ 𝑥 < 0.18.0 |
𝑥
= Vulnerable software versions
Early Detection
Affected products identified ahead of NVD analysis through intelligence sources.
| Vendor | Product | Version | Source |
|---|---|---|---|
| Red Hat | Red Hat AI Inference Server 3.2 | 1779223654 ≤ 𝑥 < * | ADP |
| Red Hat | Red Hat AI Inference Server 3.2 | 1779223651 ≤ 𝑥 < * | ADP |
| Red Hat | Red Hat AI Inference Server 3.3 | 1775680192 ≤ 𝑥 < * | ADP |
| Red Hat | Red Hat AI Inference Server 3.3 | 1775680262 ≤ 𝑥 < * | ADP |
| Red Hat | Red Hat AI Inference Server 3.3 | 1775749857 ≤ 𝑥 < * | ADP |
| Red Hat | Red Hat Enterprise Linux AI 3.3 | 1776871984 ≤ 𝑥 < * | ADP |
| Red Hat | Red Hat Enterprise Linux AI 3.3 | 1776871985 ≤ 𝑥 < * | ADP |
| Red Hat | Red Hat Enterprise Linux AI 3.3 | 1776872005 ≤ 𝑥 < * | ADP |
| Red Hat | Red Hat Enterprise Linux AI 3.3 | 1776773390 ≤ 𝑥 < * | ADP |
| Red Hat | Red Hat Enterprise Linux AI 3.3 | 1776871987 ≤ 𝑥 < * | ADP |
| Red Hat | Red Hat Enterprise Linux AI 3.3 | 1776773505 ≤ 𝑥 < * | ADP |
| Red Hat | Red Hat Enterprise Linux AI 3.3 | 1776938871 ≤ 𝑥 < * | ADP |
| Red Hat | Red Hat OpenShift AI 2.25 | 1780069069 ≤ 𝑥 < * | ADP |
| Red Hat | Red Hat OpenShift AI 2.25 | 1782397826 ≤ 𝑥 < * | ADP |
| Red Hat | Red Hat OpenShift AI 2.25 | 1783998774 ≤ 𝑥 < * | ADP |
| Red Hat | Red Hat OpenShift AI 2.25 | 1783998857 ≤ 𝑥 < * | ADP |
| Red Hat | Red Hat OpenShift AI 3.3 | 1778600187 ≤ 𝑥 < * | ADP |
| Red Hat | Red Hat OpenShift AI 3.3 | 1782996080 ≤ 𝑥 < * | ADP |
Common Weakness Enumeration
- CWE-693 - Protection Mechanism FailureThe product does not use or incorrectly uses a protection mechanism that provides sufficient defense against directed attacks against the product.
- CWE-501 - Trust Boundary ViolationThe product mixes trusted and untrusted data in the same data structure or structured message.
References