CyberRota Analysis
AI-GeneratedThe vLLM inference engine versions 0.19.0 to 0.26.0 are vulnerable due to the /v1/completions CompletionRequest.prompt field allowing an unbounded input size, which can lead to resource exhaustion, impacting CPU, memory, and async scheduling. This vulnerability poses a risk to authenticated API clients, enabling them to overwhelm system resources with a single request. Organizations utilizing affected versions of vLLM should prioritize upgrading to version 0.26.0 to mitigate potential service disruptions.
Public Exploit Signal
A public exploit, PoC, GitHub repository or Metasploit reference was detected for this CVE.
Note: these links are listed for security research and verification purposes only.
Original NVD Description
vLLM is an inference and serving engine for large language models. From 0.19.0 until 0.26.0, the /v1/completions CompletionRequest.prompt field in vllm/entrypoints/openai/completion/protocol.py accepts an unbounded list[str] or list[list[int]], prompt_to_seq() in vllm/renderers/inputs/preprocess.py and OnlineRenderer.preprocess_completion() in vllm/renderers/online_renderer.py expand every element, and vllm/entrypoints/openai/completion/serving.py creates one engine generator and response slot per prompt, allowing an authenticated API client to exhaust CPU, memory, async scheduling capacity, engine request slots, and response buffering with one request. This issue is fixed in version 0.26.0.