PoC Monitor
← All CVEs
CVE-2026-100651published 26 Sep 2026 13:23 UTCData: JSON · API

Vllm-project vllm

vllm before 0.29.0 Denial of Service via Decoder Prompt Length Bypass

vLLM before 0.29.0 fails to enforce decoder prompt-length validation on the disaggregated serving endpoint /inference/v1/generate. When the request contains a 'features' (multimodal) payload, vllm/entrypoints/serve/disagg/serving.py builds a multimodal EngineInput directly from the caller-supplied token_ids, and GenerateRequest.token_ids (vllm/entrypoints/serve/disagg/protocol.py) is not checked against model_config.max_model_len. For multimodal processors that report skip_prompt_length_check=True (for example Nemotron Parse, Whisper, and FireRedLID), InputProcessor._validate_prompt_len() returns immediately for both encoder and decoder prompts, so an overlong prompt becomes an EngineCoreRequest and reaches the worker input-batch copy into a fixed max_model_len-wide NumPy row. A client able to reach the endpoint on an affected model configuration can therefore submit an overlong token_ids list to trigger a worker failure and denial of service. Fixed in 0.29.0.

First public PoC None yetno public PoC found
Public PoCs00 on GitHub
CVSS7.1high · scored by CNA · CWE-400
CISA KEV Not listednot known exploited

Affected versions and mitigation

From the CVE record and the vendor's references.

Affected

vllm-project vllm
  • < 0.29.0

Fixed in 0.29.0

Mitigation

Upgrade vllm to 0.29.0 or later.

References

Timeline

Days relative to CVE publication
-2t=0+2+4+6+8CVE published26 Sep 2026 · t = 0

No public PoC found yet; not in CISA KEV.

PoC repositories

Found on GitHub.

No PoC repositories found yet.