Impact

This report is to highlight a vulnerability in XGrammar, a library used by the structured output feature in vLLM. The XGrammar advisory is here: GHSA-389x-67px-mjg3

The xgrammar library is the default backend used by vLLM to support structured output (a.k.a. guided decoding). Xgrammar provides a required, built-in cache for its compiled grammars stored in RAM. xgrammar is available by default through the OpenAI compatible API server with both the V0 and V1 engines.

A malicious user can send a stream of very short decoding requests with unique schemas, resulting in an addition to the cache for each request. This can result in a Denial of Service by consuming all of the system's RAM.

Note that even if vLLM was configured to use a different backend by default, it is still possible to choose xgrammar on a per-request basis using the guided_decoding_backend key of the extra_body field of the request with the V0 engine. This per-request choice is not available when using the V1 engine.

Patches

• vllm-project/vllm#16283

Workarounds

There is no way to workaround this issue in existing versions of vLLM other than preventing untrusted access to the OpenAI compatible API server.

References

• GHSA-389x-67px-mjg3

References

• GHSA-389x-67px-mjg3
• GHSA-hf3c-wxg2-49q9
• vllm-project/vllm#16283
• vllm-project/vllm@cb84e45