Medium severity4.8NVD Advisory· Published Jun 17, 2026· Updated Jul 7, 2026
CVE-2026-12491
CVE-2026-12491
Description
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.
AI Insight
LLM-synthesized narrative grounded in this CVE's description and references.
Affected packages
Versions sourced from the GitHub Security Advisory.
| Package | Affected versions | Patched versions |
|---|---|---|
vllmPyPI | >= 0.11.0, < 0.24.0 | 0.24.0 |
Affected products
5- osv-coords4 versionspkg:apk/chainguard/vllm-cuda-13.2pkg:apk/chainguard/vllm-openai-cuda-13.0pkg:apk/chainguard/vllm-openai-cuda-12.9pkg:apk/chainguard/tritonserver-backend-vllm-cuda-13.0
< 0.24.0-r0+ 3 more
- (no CPE)range: < 0.24.0-r0
- (no CPE)range: < 0.24.0-r1
- (no CPE)range: < 0.28.0-r0
- (no CPE)range: < 25.11-r12
Patches
Vulnerability mechanics
References
9- github.com/advisories/GHSA-8jr5-v98p-w75mghsaADVISORY
- nvd.nist.gov/vuln/detail/CVE-2026-12491ghsaADVISORY
- access.redhat.com/security/cve/CVE-2026-12491nvdWEB
- bugzilla.redhat.com/show_bug.cginvdWEB
- github.com/pypa/advisory-database/tree/main/vulns/vllm/PYSEC-2026-3406.yamlghsaWEB
- github.com/vllm-project/vllm/commit/cf1c90672404548aa3bc51f92c4745576a65ee26ghsaWEB
- github.com/vllm-project/vllm/pull/44974ghsaWEB
- github.com/vllm-project/vllm/security/advisories/GHSA-8jr5-v98p-w75mghsaWEB
- pypi.org/project/vllmghsaWEB
News mentions
0No linked articles in our index yet.