CVE-2026-100653
mediumpriority · track · 26EPSS 0.003 · 27 %ileDEUSEPSS · scored 2026-10-27
vLLM is an inference and serving engine for large language models. In versions from 0.22.1 through 0.28.0, the operator-supplied model revision pin (--revision / --code-revision) is not propagated to several Hugging Face artifact loads for the FunAudioChat and Tarsier2 architectures: the WhisperFeatureExtractor and speech_tokenizer PreTrainedTokenizerFast loads in vllm/model_executor/models/funaudiochat.py and the Qwen2VLConfig.from_pretrained call used by Tarsier2ProcessingInfo in vllm/model_executor/models/qwen2_vl.py. As a result, deployments pinned to a reviewed revision still resolve these behavior-affecting processor, tokenizer, and config artifacts from the repository's default revision, so a later change to the upstream default branch can alter audio preprocessing, speech tokenizer behavior, or Tarsier2 configuration without any change to the operator's configured pin. This is a supply-chain integrity and reproducibility failure for pinned deployments; it is residual to the earlier fix tracked as GHSA-3ww4-5jv9-j5gm / CVE-2026-47155 and does not constitute remote code execution or a trust_remote_code=False bypass. The issue is fixed in version 0.28.0.
cvss
Technical severity, 0–10. One row per scoring source.
base score 0–10 · higher = more severe · hover a vector to decode⚠ sources disagree · medium–high · 4.8–8.3
ssvc
A stakeholder decision — act, attend, or track — from exploitation and impact.
weaknesses
The CWE weakness classes this flaw belongs to.
affected products (4)
Products and version ranges known to be affected.
advisories (2)
Vendor and ecosystem advisories describing this.
- WID-SEC-W-2026-3327vllm: Mehrere Schwachstellen ermöglichenDeutsch2026-09-13csaf · cert_bundDE
Ein Angreifer kann mehrere Schwachstellen in vllm ausnutzen, um Sicherheitsvorkehrungen zu umgehen, Daten zu manipulieren, sensible Informationen offenzulegen oder einen Denial-of-Service-Zustand auszulösen.
- PYSEC-2026-43242026-09-26vendor · osvUS
vLLM is an inference and serving engine for large language models. In versions from 0.22.1 through 0.28.0, the operator-supplied model revision pin (--revision / --code-revision) is not propagated to several Hugging Face artifact loads for the FunAudioChat and Tarsier2 architectures: the WhisperFeatureExtractor and speech_tokenizer PreTrainedTokenizerFast loads in vllm/model_executor/models/funaudiochat.py and the Qwen2VLConfig.from_pretrained call used by Tarsier2ProcessingInfo in vllm/model_executor/models/qwen2_vl.py. As a result, deployments pinned to a reviewed revision still resolve these behavior-affecting processor, tokenizer, and config artifacts from the repository's default revision, so a later change to the upstream default branch can alter audio preprocessing, speech tokenizer behavior, or Tarsier2 configuration without any change to the operator's configured pin. This is a supply-chain integrity and reproducibility failure for pinned deployments; it is residual to the earlier fix tracked as GHSA-3ww4-5jv9-j5gm / CVE-2026-47155 and does not constitute remote code execution or a trust_remote_code=False bypass. The issue is fixed in version 0.28.0.
also known as (1)
Other names and identifiers for this entity.
sources
Which sources assert this record, and when.
references (3) nvd
External citations — patches, advisories, analysis.
- https://github.com/vllm-project/vllm/commit/d26a28ab033697f55a1414b5b0435de7cd6045b6Patch
- https://github.com/vllm-project/vllm/security/advisories/GHSA-hhv2-872h-628qExploitMitigationVendor Advisory
- https://www.vulncheck.com/advisories/vllm-0.22.1-before-0.28.0-incomplete-artifact-pin-propagationThird Party Advisory
Attackers weaponise vulnerabilities like this against internet-exposed services. See what is exposed on your perimeter with a free passive scan.
Free scan