Software
Breachwright← Back to the database
medium · 4.8Maintainer reviewedAASD-2026-0004
NVIDIA Triton model configuration handling permits information disclosure or denial of service
NVIDIA Triton Inference Server releases before r26.02 insufficiently constrained a path derived while processing an uploaded model configuration. A network caller able to supply a model configuration to an affected server could cause limited information disclosure or disrupt inference service availability.
Prerequisites
- An NVIDIA Triton Inference Server release before r26.02 is reachable
- The deployment permits the attacker to supply a model configuration to the vulnerable processing path
Potential impact
- Limited disclosure of information accessible to the Triton server process
- Denial of service affecting model inference availability
Mitigations
- Upgrade NVIDIA Triton Inference Server to r26.02 or later
- Do not expose model-management or repository-control interfaces to untrusted networks
- Accept model configurations only from authenticated, authorized deployment pipelines
- Run inference services with least privilege and isolate model repositories from unrelated sensitive data
Public sources
Research note
Not listed in the CISA Known Exploited Vulnerabilities catalog when verified on 2026-07-28.