LightLLM through 1.2.0 fails to validate image_url and audio_url parameters in multimodal endpoints, allowing unauthenticated attackers to perform server-side request forgery. Attackers can supply arbitrary URLs to fetch internal resources, with vision model processing disclosing content or error responses revealing internal network topology.
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.
Related Attack Patterns
References
- https://github.com/ModelTC/LightLLM
- https://github.com/ModelTC/LightLLM/issues/1608
- https://github.com/ModelTC/lightllm/blob/v1.2.0/lightllm/server/multimodal_params.py#L149
- https://github.com/ModelTC/lightllm/blob/v1.2.0/lightllm/utils/multimodal_utils.py#L82-L91
- https://www.vulncheck.com/advisories/lightllm-through-1.2.0-server-side-request-forgery-via-multimodal-endpoints
- https://github.com/ModelTC/LightLLM/issues/1608