CVE-2026-93592: Improper Validation of Array Index in vllm-project vllm
vLLM versions before 0.28.0 fail to validate the lower bound of token IDs in the /v1/embeddings and /pooling endpoints, allowing unauthenticated attackers to crash the engine by submitting negative token IDs. A single request with a negative token ID triggers a CUDA device-side assertion that poisons the GPU context, causing all subsequent requests to fail until the process restarts.
AI Analysis
Technical Summary
CVE-2026-93592 describes an improper validation of array index vulnerability in vLLM versions prior to 0.28.0. Specifically, the software does not validate that token IDs are non-negative in the /v1/embeddings and /pooling API endpoints. An attacker can submit a negative token ID which triggers a CUDA device-side assertion failure. This failure poisons the GPU context, causing the engine to crash and reject all further requests until the process is restarted.
Potential Impact
An unauthenticated attacker can cause a denial of service by crashing the vLLM engine through a single request with a negative token ID. The crash poisons the GPU context, making the service unavailable for all subsequent requests until a manual or automatic process restart occurs. There is no indication of data leakage or code execution from this vulnerability.
Mitigation Recommendations
Upgrade to vLLM version 0.28.0 or later where this issue is fixed. Since the vulnerability affects versions before 0.28.0, applying the official fix by updating the software is recommended. No vendor advisory or patch link was provided, so confirm the patch availability from the vendor's official channels before deployment.
CVE-2026-93592: Improper Validation of Array Index in vllm-project vllm
Description
vLLM versions before 0.28.0 fail to validate the lower bound of token IDs in the /v1/embeddings and /pooling endpoints, allowing unauthenticated attackers to crash the engine by submitting negative token IDs. A single request with a negative token ID triggers a CUDA device-side assertion that poisons the GPU context, causing all subsequent requests to fail until the process restarts.
CVSS v4.0
Score 8.7high
Affected software
vllm-project
vllm
Run on your own infrastructure? Check whether these packages are installed with threat-finder — our free open-source scanner.
AI-Powered Analysis
Machine-generated threat intelligence
Technical Analysis
CVE-2026-93592 describes an improper validation of array index vulnerability in vLLM versions prior to 0.28.0. Specifically, the software does not validate that token IDs are non-negative in the /v1/embeddings and /pooling API endpoints. An attacker can submit a negative token ID which triggers a CUDA device-side assertion failure. This failure poisons the GPU context, causing the engine to crash and reject all further requests until the process is restarted.
Potential Impact
An unauthenticated attacker can cause a denial of service by crashing the vLLM engine through a single request with a negative token ID. The crash poisons the GPU context, making the service unavailable for all subsequent requests until a manual or automatic process restart occurs. There is no indication of data leakage or code execution from this vulnerability.
Mitigation Recommendations
Upgrade to vLLM version 0.28.0 or later where this issue is fixed. Since the vulnerability affects versions before 0.28.0, applying the official fix by updating the software is recommended. No vendor advisory or patch link was provided, so confirm the patch availability from the vendor's official channels before deployment.
Technical Details
- Data Version
- 5.2
- Assigner Short Name
- VulnCheck
- Date Reserved
- 2026-09-18T10:59:45.138Z
- Cvss Version
- 4.0
- State
- PUBLISHED
Threat ID: 6aad3d5b55bf5e2cf51004ee
Added to database: 09/18/2026, 13:32:11 UTC
Last enriched: 09/18/2026, 13:47:55 UTC
Last updated: 09/18/2026, 22:35:27 UTC
Views: 8
Community Reviews
0 reviewsCrowdsource mitigation strategies, share intel context, and vote on the most helpful responses. Sign in to add your voice and help keep defenders ahead.
Want to contribute mitigation steps or threat intel context? Sign in or create an account to join the community discussion.
Actions
Updates to AI analysis require Pro Console access. Upgrade inside Console → Billing.
Need more coverage?
Upgrade to Pro Console for AI refresh and higher limits.
For incident response and remediation, OffSeq services can help resolve threats faster.
Latest Threats
Check if your credentials are on the dark web
Instant breach scanning across billions of leaked records. Free tier available.