GHSA-x8xr-mj9x-6h7wMedium· 4.8▾ SunlitDuplicate Advisory: image EXIF Rotation & PNG tRNS Transparency Not Normalized, Causing Mismatch Between Model Input and Expectations
▾ Sunlit zone — Low / medium · no exploitation signal
impact 26.4 · likelihood 0 · exploitation 0
Need a working PoC? Pro members can cast a request and our team develops one — it lands right here.
Disclosure to exploitation, from the record and what we observed since indexing it.
Disclosed via GHSA
Last analysed / modified upstream
This advisory has been withdrawn because it is a duplicate of GHSA-8jr5-v98p-w75m. This link is maintained to preserve external references.
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.
vllm >= 0.11.0, <= 0.23.0Refer to the advisory for the patched release.
Connected by shared product, vendor, weakness, or advisory.
CVE-2026-12491Medium· 4.8vLLM: image EXIF Rotation & PNG tRNS Transparency Not Normalized, Causing Mismatch Between Model Input and Expectations
GHSA-8qpc-6rxx-4pxfHigh· 7.5Duplicate Advisory: vllm p2p never connected zmq_exhaustion
GHSA-f623-hjm8-5wg5Critical· 8.1Duplicate Advisory: vLLM: Request-controlled mm_processor_kwargs.code_revision allows remote code execution
CVE-2026-115911Medium· 6.5vLLM: DoS via unbounded chat_template Jinja2 rendering exhausts the renderer thread pool
CVE-2026-115912High· 7.5vLLM: Public multimodal SHM handle forgery reaches worker-side `pickle.loads()` with insecure serialization disabled
CVE-2026-115913Medium· 6.5vLLM: Phi-4 multimodal audio encoder materializes and retains full attention matrices: a single max-length audio OOM-crashes the vLLM engine