CVSS 3.1
4.8 mediumCVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:L/A:LCVE-2026-12491
Medium (4.8). EPSS 0.2% chance of exploitation in the next 30 days. Fixed in vllm 0.24.0.
A flaw was found in vLLM, an open-source library for large language model inference. This vulnerability arises from improper handling of image metadata, specifically EXIF orientation and PNG transparency (tRNS) data, during image processing. When images are converted to RGB, transparency information may be implicitly discarded or remapped, leading to unexpected rendering of transparent pixels and distortion of input content. This can result in the model misinterpreting image content, potentially affecting the integrity of processed data.
Severity
- EPSS
- 0.2% chance of exploitation in the next 30 days; higher than 15% of CVEs (scores of 2026-10-31)
- CWE
- CWE-115
- Published
- Modified
- Record from
- the NVD, synced
Packages (OSV)
| Advisory | Ecosystem | Package | Affected | Fixed in |
|---|---|---|---|---|
| GHSA-8jr5-v98p-w75m | PyPI | vllm | >=0.11.0 <0.24.0 | 0.24.0 |
| PYSEC-2026-3406 | PyPI | vllm | >=0.11.0 <=0.23.0 | no fix published |