Summary

The /v1/chat/completions and /tokenize endpoints allow a chat_template_kwargs request parameter that is used in the code before it is properly validated against the chat template. With the right chat_template_kwargs parameters, it is possible to block processing of the API server for long periods of time, delaying all other requests

Details

In serving_engine.py, the chat_template_kwargs are unpacked into kwargs passed to chat_utils.py apply_hf_chat_template with no validation on the keys or values in that chat_template_kwargs dict. This means they can be used to override optional parameters in the apply_hf_chat_template method, such as tokenize, changing its default from False to True.

https://github.com/vllm-project/vllm/blob/2a6dc67eb520ddb9c4138d8b35ed6fe6226997fb/vllm/entrypoints/openai/serving_engine.py#L809-L814

https://github.com/vllm-project/vllm/blob/2a6dc67eb520ddb9c4138d8b35ed6fe6226997fb/vllm/entrypoints/chat_utils.py#L1602-L1610

Both serving_chat.py and serving_tokenization.py call into this _preprocess_chat method of serving_engine.py and they both pass in chat_template_kwargs.

So, a chat_template_kwargs like {"tokenize": True} makes tokenization happen as part of applying the chat template, even though that is not expected. Tokenization is a blocking operation, and with sufficiently large input can block the API server's event loop, which blocks handling of all other requests until this tokenization is complete.

This optional tokenize parameter to apply_hf_chat_template does not appear to be used, so one option would be to just hard-code that to always be False instead of allowing it to be optionally overridden by callers. A better option may be to not pass chat_template_kwargs as unpacked kwargs but instead as a dict, and only unpack them after the logic in apply_hf_chat_template that resolves the kwargs against the chat template.

Impact

Any authenticated user can cause a denial of service to a vLLM server with Chat Completion or Tokenize requests.

Fix

vllm-project/vllm#27205

References

• GHSA-69j4-grxj-j64p
• vllm-project/vllm#27205
• vllm-project/vllm@3ada34f
• https://github.com/vllm-project/vllm/blob/2a6dc67eb520ddb9c4138d8b35ed6fe6226997fb/vllm/entrypoints/chat_utils.py#L1602-L1610
• https://github.com/vllm-project/vllm/blob/2a6dc67eb520ddb9c4138d8b35ed6fe6226997fb/vllm/entrypoints/openai/serving_engine.py#L809-L814
• https://nvd.nist.gov/vuln/detail/CVE-2025-62426