CVE-2025-23268
Published: 17 September 2025
Summary
CVE-2025-23268 is a high-severity Improper Input Validation (CWE-20) vulnerability in Nvidia Triton Inference Server. Its CVSS base score is 8.0 (High).
Operationally, exploitation aligns with the MITRE ATT&CK technique Exploit Public-Facing Application (T1190); ranked at the 38.0th percentile by exploit likelihood (below the median); it is not currently listed in the CISA KEV catalog.
The strongest mitigations our analysis identified are NIST 800-53 SI-10 (Information Input Validation) and SI-2 (Flaw Remediation).
Threat & Defense at a Glance
Threat & Defense Details
Mitigating Controls (NIST 800-53 r5)AI
Directly enforces validation of inputs to the DALI backend, comprehensively addressing the improper input validation (CWE-20) that enables code execution.
Mandates timely identification, reporting, and correction of the specific flaw in NVIDIA Triton Inference Server, such as applying patches from the NVIDIA security bulletin.
Enforces least privilege to limit high-privilege (PR:H) access to the vulnerable Triton Inference Server, reducing the pool of potential attackers able to exploit over the network.
MITRE ATT&CK Enterprise TechniquesAI
Why these techniques?
RCE via improper input validation in network-accessible Triton Inference Server directly enables T1190 (Exploit Public-Facing Application).
NVD Description
NVIDIA Triton Inference Server contains a vulnerability in the DALI backend where an attacker may cause an improper input validation issue. A successful exploit of this vulnerability may lead to code execution.
Deeper analysisAI
CVE-2025-23268 affects the NVIDIA Triton Inference Server, specifically a vulnerability in the DALI backend stemming from improper input validation (CWE-20). Published on 2025-09-17, this issue allows an attacker to potentially achieve code execution upon successful exploitation. The vulnerability carries a CVSS v3.1 base score of 8.0 (AV:N/AC:H/PR:H/UI:N/S:C/C:H/I:H/A:H), indicating high severity with network accessibility but requiring high privileges and complexity.
An attacker must have high privileges (PR:H) on the target system and leverage network access (AV:N) with a high-complexity attack (AC:H), requiring no user interaction (UI:N). Successful exploitation results in scoped impact (S:C) with high effects on confidentiality, integrity, and availability (C:H/I:H/A:H), enabling arbitrary code execution on the server.
Mitigation details are provided in the NVIDIA security bulletin at https://nvidia.custhelp.com/app/answers/detail/a_id/5691.
Details
- CWE(s)