01 Record
CVE-2026-34760
vLLM: Downmix Implementation Differences as Attack Vectors Against Audio AI Models
MediumPUBLISHEDCNA: GitHub_M
CNA base score5.9 CVSS v3.xCVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:H/A:L
02 Description
vLLM: Downmix Implementation Differences as Attack Vectors Against Audio AI Models
03 Exploitation and scoring
- Exploitation
- Assessed, none found. CISA's assessment recorded neither public exploit code nor evidence of exploitation when it was made.
- SSVC decision
- ExploitationnoneCISA recorded no public exploit code and no evidence of exploitation.
- AutomatablenoAt least one step from reconnaissance to exploitation cannot be reliably automated.
- Technical ImpactpartialSuccessful exploitation gives limited control of the vulnerable component.
- EPSS probability
- 0.26% probability of exploitation activity in the next 30 days.
- EPSS percentile
- Equal to or higher than the score of 18.6% of scored records. A rank, not a probability.
- NVD base score
- 7.1 (CVSS v3.x), scored by the NVD independently of the CNA.
- NVD CVSS vector
- CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:H/A:L
- NVD analysis status
- Analyzed. NVD last modified this record on 2026-07-24.
- Red Hat severity
- Moderate. Red Hat's rating for its own products, not a CVSS rating.
Exploit-prediction scores from FIRST. Independent scoring and analysis status from the NVD. Vendor severity from Red Hat. Exploitation assessment and catalog membership from CISA.
04 Metadata
- Published
- 2026-04-02 18:59Z6 MO AGO
- Last updated
- 2026-04-03 14:42Z6 MO AGO
- Reserved
- 2026-03-30
- Assigning CNA
- GitHub_M
- Record state
- PUBLISHED
- Severity
- Medium (CVSS 5.9)
- CNA CVSS vector
- CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:H/A:L
- Weaknesses
- Data version
- 5.2
- Document digest
- 6f2233c75ec548803d5574b6d04323344ae554b78a45cb447a28feae381afe52
05 Affected products
VendorProductVersionsPlatforms
06 References
- github.comhttps://github.com/vllm-project/vllm/security/advisories/GHSA-6c4r-fmh3-7rh8
- github.comhttps://github.com/vllm-project/vllm/pull/37058
- github.comhttps://github.com/vllm-project/vllm/commit/c7f98b4d0a63b32ed939e2b6dfaa8a626e9b46c4
- github.comhttps://github.com/vllm-project/vllm/releases/tag/v0.18.0