---
cve: "CVE-2026-100653"
severity: "HIGH"
cvss: 8.3
epss: "0.3%"
vendor: "vllm-project"
kev: false
exploited: false
published: "2026-09-26 14:16:47"
tags: [cve, security, high]
source: tsecurity.de CVE-Dossier
exported: "2026-10-02T22:22:10+02:00"
---

# CVE-2026-100653

> 8.3 HIGH · 🧪 PoC

## Beschreibung

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.

## CNA-Record (Kanon, cvelistV5)

- CNA: **VulnCheck**
- State: PUBLISHED
- Stand: 2026-10-01 11:06:45
- CNA-CVSS: **8.3** (`CVSS:4.0/AV:N/AC:H/AT:P/PR:N/UI:N/VC:L/VI:H/VA:N/SC:N/SI:N/SA:N`)

## CISA-Anreicherung (vulnrichment/ADP)

- CISA-SSVC (Coordinator, v2.0.3):
  - Exploitation: **poc**
  - Automatable: **no**
  - Technical Impact: **partial**

## CVSS-Vektor

```
CVSS:4.0/AV:N/AC:H/AT:P/PR:N/UI:N/VC:L/VI:H/VA:N/SC:N/SI:N/SA:N
```

| Metrik | Wert | Bewertung |
|---|---|---|
| AV Angriffsvektor | Netzwerk | bad |
| AC Komplexität | Hoch | good |
| PR Privilegien | Keine | bad |
| UI Interaktion | Keine | bad |

## Patch verfügbar (OSV)

- 2cf0a6915ce544dc493a0990f2ea38d81601128a (Commit)
- d26a28ab033697f55a1414b5b0435de7cd6045b6 (Commit)

## BSI-Hinweise (deutsch)

- [vllm: Mehrere Schwachstellen ermöglichen](https://wid.cert-bund.de/portal/wid/sicherheitshinweis/WID-SEC-W-2026-3327) — _BSI-Einstufung: mittel_
  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.

## Referenzen

- <https://github.com/vllm-project/vllm/security/advisories/GHSA-hhv2-872h-628q>
- <https://github.com/vllm-project/vllm/commit/d26a28ab033697f55a1414b5b0435de7cd6045b6>
- <https://www.vulncheck.com/advisories/vllm-0.22.1-before-0.28.0-incomplete-artifact-pin-propagation>

---
_Exportiert aus dem [tsecurity.de CVE-Dossier](https://tsecurity.de/cve?cve=CVE-2026-100653) · Datenquellen: EUVD (ENISA), NVD, OSV, CISA KEV, FIRST EPSS, Exploit-DB, BSI WID (CSAF)_