Disclosure summary
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.
Source-reported weakness categories
CWE-400
Source-specific records & product guidance
Sources retain their own attribution and scoring. Follow the original record to confirm affected versions, fixed releases, and configuration conditions.
NIST National Vulnerability Database · NVD-CVE-2026-100651
Open original source · Updated Oct 06, 2026
Only CPE matches marked vulnerable=true are indexed. AND/OR platform conditions must be checked in the original NVD record.
| Vendor | Product / association | Version / bounds |
|---|---|---|
| vllm | vllm | * {"versionEndExcluding":"0.29.0"} |
Original records & references
- NIST NVD record
- CVE Program record
- github.com — Exploit, Vendor Advisory
- www.vulncheck.com — Third Party Advisory
- github.com — Exploit, Vendor Advisory
PUBLISHED 2026-09-26T10:16:47-04:00
MODIFIED 2026-10-06T15:30:25-04:00
INGESTED 2026-10-10T20:45:41-04:00