vLLM is an inference and serving engine for large language models. Prior to 0.26.0, the MiMoV2OmniMultiModalProcessor in vllm/transformers_utils/processors/mimo_v2_omni.py passes attacker-controlled image and audio strings through _fetch_image, requests.get, and Image.open instead of MediaConnector, bypassing allowed_media_domains and allowed_local_media_path protections and allowing server-side requests and reads of arbitrary files accessible to the vLLM process. This issue is fixed in version 0.26.0.
Weakness
The web server receives a URL or similar request from an upstream component and retrieves the contents of this URL, but it does not sufficiently ensure that the request is being sent to the expected destination.
Affected Software
| Name | Vendor | Start Version | End Version |
|---|---|---|---|
| Vllm | Vllm | * | 0.26.0 (excluding) |
Related Attack Patterns
References
- https://github.com/vllm-project/vllm/commit/54503ecec0f3ac31e5ecfc5f28652e4cc42307b5
- https://github.com/vllm-project/vllm/pull/43117
- https://github.com/vllm-project/vllm/releases/tag/v0.26.0
- https://github.com/vllm-project/vllm/security/advisories/GHSA-4hhp-h66f-j5j7
- https://github.com/vllm-project/vllm/security/advisories/GHSA-4hhp-h66f-j5j7