CVE-2026-94623
vLLM through 0.29.0 contains a denial of service vulnerability in the NIXL connector's prefix caching implementation that fails to properly
CVSS
7.5
High
EPSS
—
KEV
—
Exploit Today
—
0-100
Published: Sep 21, 2026 · Last modified: Sep 21, 2026 · CWE-617
Not enough EPSS history yet.
vLLM through 0.29.0 contains a denial of service vulnerability in the NIXL connector's prefix caching implementation that fails to properly validate block counts across multi-prompt completion requests in prefill/decode disaggregated deployments. Attackers can trigger an assertion failure in NixlBaseConnectorWorker._apply_prefix_caching by submitting completion requests with multiple prompts of varying lengths, causing the decode worker to terminate and become unavailable until restarted.
- github.comhttps://github.com/vllm-project/vllm
- github.comhttps://github.com/vllm-project/vllm/blob/v0.29.0/vllm/distributed/kv_transfer/kv_connector/v1/nixl/base_worker.py#L2642-L2646
- github.comhttps://github.com/vllm-project/vllm/pull/51505
- www.vulncheck.comhttps://www.vulncheck.com/advisories/vllm-through-0.29.0-denial-of-service-via-nixl-multi-prompt-assertion-failure