CVE-2026-115916

Published: Ott 26, 2026 Last Modified: Ott 26, 2026

ExploitDB:

Other exploit source:

Google Dorks:

MEDIUM 6,5

Source: [email protected]

Attack Vector: network

Attack Complexity: low

Privileges Required: low

User Interaction: none

Scope: unchanged

Confidentiality: none

Integrity: none

Availability: high

Description

vLLM is an inference and serving engine for large language models. Prior to 0.32.0, vllm/tokenizers/hf.py computes max_token_id with an off-by-one upper bound, allowing a pre-tokenized prompt ID equal to the vocabulary size to pass validation. For a tensor-parallel size of one and an exact multiple-of-64 vocabulary, VocabParallelEmbedding.forward does not mask that value before F.embedding uses it, causing a CUDA device-side assertion and terminating the engine. This issue is fixed in version 0.32.0.

129

Improper Validation of Array Index

Draft

Abstraction: Variant Structure: Simple

Common Consequences

Security Scopes Affected:

Integrity Availability Confidentiality

Potential Impacts:

Dos: Crash, Exit, Or Restart Modify Memory Read Memory Execute Unauthorized Code Or Commands

Applicable Platforms

Languages: C, C++, Not Language-Specific

Likelihood of Exploit: High

View CWE Details

193

Off-by-one Error

Draft

Abstraction: Base Structure: Simple

Common Consequences

Security Scopes Affected:

Availability Integrity Confidentiality Access Control

Potential Impacts:

Dos: Crash, Exit, Or Restart Dos: Resource Consumption (Cpu) Dos: Resource Consumption (Memory) Dos: Instability Modify Memory Execute Unauthorized Code Or Commands Bypass Protection Mechanism