What is the CVSS score of CVE-2026-24181?

CVE-2026-24181 has a CVSS 3.1 base score of 7.3 (High). CVSS vector: CVSS:3.1/AV:L/AC:L/PR:L/UI:R/S:U/C:H/I:H/A:H. EPSS exploitation probability: 0.0%.

Which versions of NVIDIA DALI are affected by CVE-2026-24181?

NVIDIA DALI (Data Loading Library) contains a local code execution vulnerability (CVE-2026-24181, CVSS 7.3) that enables low-privilege users to corrupt data and access sensitive information when…

How to fix or mitigate CVE-2026-24181?

24 hours: Inventory all systems, applications, and teams using NVIDIA DALI; document data sources and operational criticality. 7 days: Implement input validation and source-of-truth controls for all DALI inputs; run DALI processes with minimum necessary privileges; restrict file access and script execution permissions. 30 days: Monitor NVIDIA security advisories for patch release; establish escalation procedures if patch timeline extends beyond 60 days.

NVIDIA DALI CVE-2026-24181

EUVD-2026-35498 HIGH

Improper Validation of Array Index (CWE-129)

2026-06-09 nvidia

GHSA-g7x7-xpr8-2r5h

RCE Denial Of Service Information Disclosure Nvidia Dali

7.3

CVSS 3.1 · Vendor: nvidia

Severity by source

Vendor (nvidia) PRIMARY

7.3 HIGH

AV:L/AC:L/PR:L/UI:R/S:U/C:H/I:H/A:H

Primary rating from Vendor (nvidia) · only source for this CVE.

CVSS VectorVendor: nvidia

Attack Vector

Local

Attack Complexity

Low

Privileges Required

Low

User Interaction

Required

Scope

Unchanged

Confidentiality

High

Integrity

High

Availability

High

Lifecycle Timeline

Analysis Generated

Jun 09, 2026 - 16:49 vuln.today

DescriptionCVE.org

NVIDIA DALI contains a vulnerability in a component where an attacker could cause an improper index validation. A successful exploit of this vulnerability might lead to code execution, data tampering, denial of service, and information disclosure.

AnalysisAI

Local code execution and data tampering in NVIDIA DALI (Data Loading Library) is possible when a low-privileged user is tricked into processing attacker-controlled input through a component that performs improper index validation (CWE-129). The CVSS 7.3 vector (AV:L/AC:L/PR:L/UI:R) indicates local access, low privileges, and user interaction are required, with high impact across confidentiality, integrity, and availability. No public exploit identified at time of analysis and the issue is not listed in CISA KEV.

Technical ContextAI

NVIDIA DALI (Data Loading and Augmentation Library) is a GPU-accelerated library used in deep-learning pipelines (PyTorch, TensorFlow, MXNet) to perform data loading and preprocessing for training and inference workloads. The CPE cpe:2.3:a:nvidia:dali:*:*:*:*:*:*:*:* indicates all versions are presently in scope pending vendor clarification. CWE-129 (Improper Validation of Array Index) describes a class of memory-safety bugs where untrusted input is used to compute an array offset without proper bounds checking, allowing out-of-bounds reads or writes that, in a native C++/CUDA library like DALI, can corrupt heap metadata or adjacent objects and pivot to arbitrary code execution within the host process.

RemediationAI

Consult NVIDIA security advisory https://nvidia.custhelp.com/app/answers/detail/a_id/5814 for the fixed DALI release and upgrade pip/conda installations (e.g., nvidia-dali-cudaXXX) to that version once published - exact fix version is not stated in the input data so cite the advisory directly rather than guessing. As compensating controls until patched, restrict who can submit DALI pipeline definitions or supply input data to trusted users only (the bug requires PR:L plus UI:R), run training/preprocessing workloads in isolated containers or VMs with non-root service accounts so that an out-of-bounds write cannot escalate to host compromise, and avoid processing untrusted user-uploaded datasets through shared DALI workers; the trade-off of isolation is added orchestration complexity and some GPU-passthrough overhead, while restricting input sources may break multi-tenant ML-as-a-service workflows.