Improper Validation of Array Index Affecting vllm package, versions [,0.29.0)
Threat Intelligence
The probability is the direct output of the EPSS model, and conveys an overall sense of the threat of exploitation in the wild. The percentile measures the EPSS probability relative to all known EPSS scores. Note: This data is updated daily, relying on the latest available EPSS model version. Check out the EPSS documentation for more details.
Do your applications use this vulnerable package?
In a few clicks we can analyze your entire application and see what components are vulnerable in your application, and suggest you quick fixes.
Test your applications- Snyk IDSNYK-PYTHON-VLLM-20158364
- published27 Sept 2026
- disclosed26 Sept 2026
- creditUnknown
Introduced: 26 Sep 2026
NewCVE-2026-100654 (opens in a new tab)How to fix?
Upgrade vllm to version 0.29.0 or higher.
Overview
vllm is an A high-throughput and memory-efficient inference and serving engine for LLMs
Affected versions of this package are vulnerable to Improper Validation of Array Index via unvalidated stop_token_ids on the OpenAI-compatible POST /v1/completions and POST /v1/chat/completions endpoints. SamplingParams in vllm/sampling_params.py validates only that each value in stop_token_ids is an integer, not that it falls within the model vocabulary range. When min_tokens > 0, the stop token ids are used as logits indices in vllm/v1/sample/logits_processor/builtin.py via a CUDA index_put_ operation. An out-of-range token id triggers a device-side assertion, putting EngineCore into a fatal state and causing all subsequent requests to fail until the service is restarted.