CVE-2026-93436

EUVD-2026-82534
vLLM through 0.29.0 fails to properly clean up decode-side metadata for rejected inference requests in prefill/decode disaggregated deployments. Remote attackers can submit requests with max_tokens=0 to exhaust decode-worker memory without bound until the worker restarts.
ProviderTypeBase ScoreAtk. VectorAtk. ComplexityPriv. RequiredVector
NISTPrimary
7.5 HIGH
NETWORK
LOW
NONE
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H
Base Score
CVSS 3.x
EPSS Score
Percentile: 55.14%
Affected Products (NVD)
VendorProductVersion
vllmvllm
𝑥
≤ 0.29.0
𝑥
= Vulnerable software versions
Common Weakness Enumeration
References
https://github.com/vllm-project/vllm
https://github.com/vllm-project/vllm/blob/v0.29.0/vllm/distributed/kv_transfer/kv_connector/v1/nixl/push_worker.py#L162-L181
https://github.com/vllm-project/vllm/pull/55677
https://www.vulncheck.com/advisories/vllm-through-0.29.0-memory-exhaustion-via-rejected-requests