CVE-2026-27893
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.
Debolezza
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.
Prodotti interessati
- vllm vllm
Riferimenti
- https://github.com/vllm-project/vllm/commit/00bd08edeee5dd4d4c13277c0114a464011acf72
- https://github.com/vllm-project/vllm/pull/36192
- https://github.com/vllm-project/vllm/security/advisories/GHSA-7972-pg2x-xr59
- https://access.redhat.com/errata/RHSA-2026:10140
- https://access.redhat.com/errata/RHSA-2026:10141
- https://access.redhat.com/errata/RHSA-2026:19712
Trova il bug prima di un attaccante.
Accedi con GitHub e avvia il tuo primo audit in meno di un minuto. Il piano gratuito non richiede carta di credito.