Unauthenticated service interruption via unchecked grid indices
Published May 12, 2026 · Updated May 12, 2026
Token injection in vLLM 0.6.1 and later before 0.20.0 allows remote attackers to interrupt VLM service with text-only prompts. The _vl_get_input_positions_tensor function treats spelled-out vision placeholders as control tokens, then indexes empty image or video grids without checking their bounds. Exploitation requires a served vision-language model that accepts raw prompt text; one request raises an unhandled IndexError and can terminate a worker or reduce capacity.
Summary
What happened
Token injection in vLLM 0.6.1 and later before 0.20.0 allows remote attackers to interrupt VLM service with text-only prompts. The _vl_get_input_positions_tensor function treats spelled-out vision placeholders as control tokens, then indexes empty image or video grids without checking their bounds. Exploitation requires a served vision-language model that accepts raw prompt text; one request raises an unhandled IndexError and can terminate a worker or reduce capacity.
The record
- CVE
- CVE-2026-44222
- Published
- May 12, 2026
- Updated
- May 12, 2026
- Vendor
- vLLM Project
- Product
- vLLM
- Classifications
- CWE-129, CAPEC-100, T1499.004
- Attack vector
- network
- Privileges
- unauthenticated
Timeline
How it unfolded
- May 12, 2026CVE publishedPublication date reported by the CVE source.
- May 12, 2026Record updatedLatest update available in the CVE record.
Exploitability
Present is not the same as exploitable
Compare your product and version with the public record. A matching version still requires validation against your environment.
Is a vulnerable build present?
Compare these published version ranges with your installed build and any vendor patches.
- Affected versionversion=>= 0.6.1, < 0.20.0
What conditions does exploitation require?
What is affected?
Attacks
What attackers are doing with it
Daily unique IPs observed by Shadowserver honeypots for known exploited vulnerabilities (KEVs). Missing observations do not establish an absence of attacks.
Weakness, pattern, technique
Public exploit references
- GHSA-hpv8-x276-m59f text-only request proof of conceptproof of concept · demonstrated
Labels summarize the accepted research assessment. They do not indicate a test against your environment.
Technologies
Your stack
See the directory against your own environment.
Your stack
Check the software in your environment
Book a demo to see how Hinoki identifies affected software and validates exploitability in your environment.
Book a demo