CVE-2026-100654

Published: Set 26, 2026 Last Modified: Set 28, 2026
ExploitDB:
Other exploit source:
Google Dorks:
HIGH 7,1
Attack Vector: network
Attack Complexity: low
Privileges Required: low
User Interaction: none
Confidentiality: N/A
Integrity: N/A
Availability: N/A
MEDIUM 6,5
Attack Vector: network
Attack Complexity: low
Privileges Required: low
User Interaction: none
Scope: unchanged
Confidentiality: none
Integrity: none
Availability: high

Description

AI Translation Available

vLLM before 0.29.0 accepts user-controlled stop_token_ids on the OpenAI-compatible POST /v1/completions and POST /v1/chat/completions endpoints but validates only that the values are integers, not that each token id is within the model vocabulary/logits range. When min_tokens > 0, the stop token ids are used as logits indices to suppress stop tokens, so an out-of-range id reaches a CUDA indexing operation (index_put_) and triggers a device-side assertion. An authenticated API user can send a single malformed completion request that returns 500 Internal Server Error and puts EngineCore into a fatal state, causing subsequent requests to fail until the service is restarted (denial of service).

129

Improper Validation of Array Index

Draft
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
View CWE Details
https://github.com/vllm-project/vllm/security/advisories/GHSA-v5gm-qgmv-gc6c
https://www.vulncheck.com/advisories/vllm-before-0.29.0-denial-of-service-via-o…