{"id":"CVE-2026-100654","published":"2026-09-26T14:16:48.100","lastModified":"2026-09-28T20:47:20.760","description":"vLLM before 0.29.0 accepts user-controlled stop_token_ids on the OpenAI-compatible POST /v1/completions and POST /v1/chat/completions endpoints but validates only that the values are integers, not that each token id is within the model vocabulary/logits range. When min_tokens > 0, the stop token ids are used as logits indices to suppress stop tokens, so an out-of-range id reaches a CUDA indexing operation (index_put_) and triggers a device-side assertion. An authenticated API user can send a single malformed completion request that returns 500 Internal Server Error and puts EngineCore into a fatal state, causing subsequent requests to fail until the service is restarted (denial of service).","cvssScore":6.5,"cvssSeverity":"MEDIUM","cvssVector":"CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H","cwes":["CWE-129"],"vendors":[],"products":[],"references":[{"url":"https://github.com/vllm-project/vllm/security/advisories/GHSA-v5gm-qgmv-gc6c","tags":[]},{"url":"https://www.vulncheck.com/advisories/vllm-before-0.29.0-denial-of-service-via-out-of-range-stop-token-ids","tags":[]}],"exploitRefs":[{"url":"https://github.com/vllm-project/vllm/security/advisories/GHSA-v5gm-qgmv-gc6c","tags":[]}],"hasPoc":true,"ai":null}