vLLM: Downmix Implementation Differences as Attack Vectors Against Audio AI Models (vllm-project vllm)
At a glance
- Source
- NVD
- EPSS
- EPSS 0.27 % · percentile 19 ()
- State
- PUBLISHED
- Published
- Updated
- Assigner
- GitHub_M
- NVD status
- Analyzed
Description
vLLM is an inference and serving engine for large language models (LLMs). From version 0.5.5 to before version 0.18.0, Librosa defaults to using numpy.mean for mono downmixing (to_mono), while the international standard ITU-R BS.775-4 specifies a weighted downmixing algorithm. This discrepancy results in inconsistency between audio heard by humans (e.g., through headphones/regular speakers) and audio processed by AI models (Which infra via Librosa, such as vllm, transformer). This issue has been patched in version 0.18.0.
Weaknesses (CWE)
- CWE-20 — Improper Input Validation
References
Scores
| Version | Score | Vector | Scored by |
|---|---|---|---|
| CVSS 3.1 | MEDIUM 5.9 | CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:H/A:L | GitHub_M |
| CVSS 3.1 | MEDIUM 5.9 | CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:H/A:L | NVD · security-advisories@github.com (Secondary) |
| CVSS 3.1 | HIGH 7.1 | CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:H/A:L | NVD · nvd@nist.gov (Primary) |
Advisories
- GHSA-6c4r-fmh3-7rh8 — vLLM: Processing differential in multi-channel audio downmixing enables hidden-input/moderation bypass for audio models
- PYSEC-2026-2299 — vLLM is an inference and serving engine for large language models (LLMs). From version 0.5.5 to before version 0.18.0, Librosa defaults to using numpy.mean for
European references
- ENISA EUVD · EUVD-2026-18522
- GCVE · GCVE-0-2026-34760
Affected configurations (NVD CPE)
cpe:2.3:a:vllm:vllm:*:*:*:*:*:*:*:*≥ 0.5.5 < 0.18.0
Affected products
| Product | Versions |
|---|---|
| vllm-project · vllm | affected >= 0.5.5, < 0.18.0 |
Sources
| Source | Licence |
|---|---|
| CVE List (CVEProject cvelistV5) | CVE Program Terms of Use |
| EPSS (FIRST.org) | FIRST — attribution requested |
| ENISA EUVD | ENISA — reproduction authorised, source acknowledged |
| NVD (NIST) via fkie-cad/nvd-json-data-feeds | Public domain — this product uses data from the NVD API but is not endorsed or certified by the NVD |