CVE-2026-115913
Tenable Plugin
Checking the public Tenable catalog…
AI Threat Summary
CVE-2026-115913 is a moderate vulnerability with a CVSS score of 6.5 affecting the target software. Threat hunters should monitor for exploit attempts and apply vendor patches immediately.
Description
vLLM is an inference and serving engine for large language models. Prior to 0.32.0, the Phi-4 multimodal audio encoder in phi4mm_audio.py and phi4mm_utils.py uses eager attention by default, allowing a remote client to submit one legal long audio clip that causes MultiHeadedAttention.forward_attention to materialize and retain full attention matrices in self.attn. The scaled dot-product attention branch does not materialize these matrices. The resulting quadratic GPU-memory use can cause an out-of-memory condition and terminate EngineCore for all users. This issue is fixed in version 0.32.0.
Attack Chain
Attacker scans for vulnerable versions using version banners or PoC scanners
Attacker triggers CVE-2026-115913 vulnerability with a crafted request or payload
Vulnerability exploited to perform unauthorized actions on the target system
Depending on vulnerability type, attacker achieves data access, service disruption, or code execution
MITRE ATT&CK Techniques
Indicators of Compromise
| Type | Value | Description | |
|---|---|---|---|
| user_agent | Mozilla/5.0 (compatible; exploit/1.0) |
Generic exploit scanner user agent pattern | |
| url | /api/vuln-endpoint?payload=<malicious> |
Exploit attempt against application endpoint — monitor access logs for anomalous params |
Remediation
Quick Facts
| CVE ID | CVE-2026-115913 |
| CVSS Score | 6.5 |
| Vector | CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H |
| Published | 2026-10-26 |
| Exploited | No |
| CISA KEV | No |
| PoC | Not public |