Server crash via unvalidated sparse tensor indices
Published Jun 20, 2026 · Updated Jun 20, 2026
Improper input validation in vLLM 0.10.2 and later before 0.13.0 allows remote authenticated users to crash the server. Multimodal embedding processing decodes attacker-supplied sparse tensors without enabling PyTorch invariant checks, so negative or out-of-bounds indices reach dense conversion and can cause resource exhaustion or out-of-bounds writes. The prompt-embeds feature must be enabled; it is disabled by default, and the published advisory provides no proof of code execution.
Summary
What happened
Improper input validation in vLLM 0.10.2 and later before 0.13.0 allows remote authenticated users to crash the server. Multimodal embedding processing decodes attacker-supplied sparse tensors without enabling PyTorch invariant checks, so negative or out-of-bounds indices reach dense conversion and can cause resource exhaustion or out-of-bounds writes. The prompt-embeds feature must be enabled; it is disabled by default, and the published advisory provides no proof of code execution.
The record
- CVE
- CVE-2026-56340
- Published
- Jun 20, 2026
- Updated
- Jun 20, 2026
- Vendor
- vLLM Project
- Product
- vLLM
- Classifications
- CWE-20, CWE-787, CWE-502, CWE-123, T1499.004
- Attack vector
- network
- Privileges
- authenticated
Timeline
How it unfolded
- Jun 20, 2026CVE publishedPublication date reported by the CVE source.
- Jun 20, 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=0.10.2 <0.13.0
What conditions does exploitation require?
What is affected?
Published CVSS scores
CVSS describes severity. EPSS estimates exploitation probability.
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
No public exploit references are available in this record.
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