CVE-2026-115921
Awaiting Analysis Awaiting Analysis - Queue

ONNX Shape Inference Null Pointer Dereference

Publication date: 2026-10-26

Last updated on: 2026-10-26

Assigner: GitHub, Inc.

Description
Open Neural Network Exchange (ONNX) is an open standard for machine learning interoperability. From 1.4.0 until 1.22.0, ScanInferenceFunction, ScanInferenceFunction_opset8, and ScanInferenceFunction_opset9 do not safely validate the required num_scan_inputs attribute. A missing attribute is dereferenced as a null pointer, while a negative value is converted to size_t and can reach an exceptional oversized allocation path. An untrusted model can therefore terminate or disrupt shape inference, including a negative-value model that can pass onnx.checker.check_model(). This issue is fixed in version 1.22.0.

CVSS Scores

EPSS Scores

Probability:
Percentile:

Meta Information

Published
2026-10-26
Last Modified
2026-10-26
Generated
2026-10-26
AI Q&A
2026-10-26
EPSS Evaluated
N/A
NVD
CVE-2026-115921
EUVD
EUVD-2026-110046

Affected Vendors & Products

Showing 1 associated CPE
Vendor Product Version / Range
onnx onnx >= 1.4.0, < 1.22.0

Helpful Resources

Exploitability

CWE
CWE Icon
KEV
KEV Icon
CWE ID Description
CWE-476 The product dereferences a pointer that it expects to be valid but is NULL.
CWE-195 The product uses a signed primitive and performs a cast to an unsigned primitive, which can produce an unexpected value if the value of the signed primitive can not be represented using an unsigned primitive.

Attack-Flow Graph

AI Quick Actions

Instant insights powered by AI
Executive Summary

This vulnerability in ONNX (Open Neural Network Exchange) involves a null pointer dereference and signed-to-unsigned conversion error in the ScanInferenceFunction. When the required attribute 'num_scan_inputs' is missing, the code crashes by dereferencing a null pointer. If the attribute is negative, converting it to an unsigned size_t causes a very large value, leading to memory allocation errors and potential denial of service.

Detection Guidance

To detect this vulnerability, check the ONNX library version installed on your system. If it is between 1.4.0 and 1.21.0, the system is vulnerable. Use commands like 'pip show onnx' or 'conda list onnx' to check the version.

Impact Analysis

The vulnerability can cause applications to terminate unexpectedly or experience denial of service when processing untrusted models. It may disrupt shape inference, leading to crashes or forced exceptional allocation paths. Confidentiality and integrity are not affected, but availability is compromised.

Mitigation Strategies

Upgrade the ONNX library to version 1.22.0 or later. This version includes the fix for the null pointer dereference and signed-to-unsigned conversion issues in ScanInferenceFunction.

Chat Assistant

Ask questions about this CVE
Hi! I’m here to help you understand CVE-2026-115921. Ask me anything about the vulnerability, its impact, or mitigation strategies.
0/70

EPSS Chart