CVE-2026-93989

Published: Sep 19, 2026Modified: Sep 28, 2026

2.3

Vector

CVSS:4.0/AV:N/AC:H/AT:P/PR:L/UI:N/VC:N/VI:L/VA:N/SC:N/SI:N/SA:N/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

Show more

CVSS:4.0/AV:N/AC:H/AT:P/PR:L/UI:N/VC:N/VI:L/VA:N/SC:N/SI:N/SA:N/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:XShow less

Source: disclosure@vulncheck.com (Secondary)

Description

vLLM through 0.29.0 fails to properly validate bad_words token indices against the model's generation output width in SamplingParams.update_from_tokenizer(). Attackers can supply out-of-bounds token indices that corrupt logits memory of concurrent requests, causing different in-flight HTTP requests to return incorrect tokens.

Affected (1)

Products: Vllm: Vllm

1 product

Configuration A

1 vulnerable

Vulnerable Software	Affected Versions
Vllm Vllm	Up to 0.29.0

Related CWEs

Improper Validation of Array Index

The product uses untrusted input when calculating or using an array index, but the product does not validate or incorrectly validates the index to ensure the index references a valid position within the array.

References (5)

https://github.com/vllm-project/vllm

Source: disclosure@vulncheck.com

Product

https://github.com/vllm-project/vllm/blob/98dff2a81d747d1dba01a47f939f48c3526d4206/vllm/sampling_params.py#L694-L753

Source: disclosure@vulncheck.com

Product

https://github.com/vllm-project/vllm/blob/98dff2a81d747d1dba01a47f939f48c3526d4206/vllm/v1/worker/gpu/sample/bad_words.py

Source: disclosure@vulncheck.com

Product

https://github.com/vllm-project/vllm/pull/48824

Source: disclosure@vulncheck.com

Issue TrackingPatch

https://www.vulncheck.com/advisories/vllm-through-0.29.0-cross-request-logits-corruption-via-bad-words

Source: disclosure@vulncheck.com

Third Party Advisory

Timeline

No history available yet.