PYSEC-2026-2297
En bref
- Publiée
- Mise à jour
- CVE
- CVE-2026-27893
- Alias
- GHSA-7972-pg2x-xr59
- CVSS_V3
- CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H
Description
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.
Paquets affectés
| Paquet | Versions |
|---|---|
| PyPI · vllm | introduced 0.10.1, fixed 0.18.0 |
Références
- https://access.redhat.com/security/cve/CVE-2026-27893 WEB
- https://security.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-27893.json WEB
- https://access.redhat.com/errata/RHSA-2026:10140 ADVISORY
- https://access.redhat.com/errata/RHSA-2026:10141 ADVISORY
- https://access.redhat.com/errata/RHSA-2026:19712 ADVISORY
- https://access.redhat.com/errata/RHSA-2026:19724 ADVISORY
- https://access.redhat.com/errata/RHSA-2026:19725 ADVISORY
- https://access.redhat.com/errata/RHSA-2026:24977 ADVISORY
- https://access.redhat.com/errata/RHSA-2026:37275 ADVISORY
- https://access.redhat.com/errata/RHSA-2026:8746 ADVISORY
- https://access.redhat.com/errata/RHSA-2026:8747 ADVISORY
- https://access.redhat.com/errata/RHSA-2026:8748 ADVISORY
- https://github.com/vllm-project/vllm/security/advisories/GHSA-7972-pg2x-xr59 ADVISORY
- https://bugzilla.redhat.com/show_bug.cgi?id=2452055 REPORT
- https://github.com/vllm-project/vllm/pull/36192 REPORT
- https://github.com/vllm-project/vllm/commit/00bd08edeee5dd4d4c13277c0114a464011acf72 FIX
Sources
| Source | Licence |
|---|---|
| OSV.dev | CC-BY 4.0 — PyPA Advisory Database |