vLLM from 0.8.5 before 0.31.0 has a denial of service flaw. LogprobsTensors.empty_cpu() in vllm/v1/outputs.py allocates an uninitialized int32 tensor, and the V1 model runner leaves prompt positions reported as computed by a KV connector unwritten. A client sharing the engine can send prompt_logprobs (or echo with logprobs) against a connector-cached prefix, making tokenizer decode raise OverflowError, which crashes EngineCore and drops all co-tenant requests.
Fix
Patch, vendor advisory and mitigation references on NVD 1
- https://github.com/vllm-project/vllm/security/advisories/GHSA-rhcx-5729-88vg Mitigation, Vendor Advisory
Every line above is quoted from NVD, CISA, Microsoft, the vendor's own advisory or DISA; nothing is generated.
Change history 8
- Cpes+1 CPE criteria (now 1)NVD modified 2026-10-28 15:40
- StatusPublished → AnalyzedNVD modified 2026-10-28 15:40
- StatusAwaiting Analysis → PublishedNVD modified 2026-10-26 19:48
- StatusPublished → Awaiting AnalysisNVD modified 2026-10-26 14:03
- StatusReceived → PublishedNVD modified 2026-10-25 13:20
- StatusPublished → ReceivedNVD modified 2026-10-25 14:16
- CVSS8.3 (v4.0, CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:H) → 8.3 (v4.0, CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:H/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X)NVD modified 2026-10-25 14:16
- First seenCVSS 8.3 · vllm-projectNVD modified 2026-10-25 13:20
Affected (CPE criteria) 1
cpe:2.3:a:vllm:vllm:*:*:*:*:*:*:*:*
References 2
- https://github.com/vllm-project/vllm/security/advisories/GHSA-rhcx-5729-88vg Mitigation, Vendor Advisory
- https://www.vulncheck.com/advisories/vllm-from-0.8.5-before-0.31.0-cross-tenant-dos-via-uninitialized-logprobstensors Third Party Advisory