Live › vulnerability › CVE-2026-115919
Go to dashboardONNX's shape inference for RNN, LSTM and GRU nodes reads past a protobuf shape when the first input has fewer than three dimensions. Processing an untrusted model can terminate the application that calls infer_shapes().
Summary written from the sources below. Check the originals before acting.
First seen by InfraStack 2026-10-26 18:31 UTC. All times on this page are UTC.
Applications that load ONNX models supplied by someone else and run shape inference on them, such as model-serving or model-conversion pipelines that accept uploads. Teams that only run models they built and trust are less exposed.
No product ranges from the sources yet.
pip show onnx and confirm the installed version is below 1.22.0onnx.shape_inference.infer_shapes run on models from outside your organizationonnx Python package to 1.22.0 or laterinfer_shapesEach row is the first time that source carried this signal. The analyst’s rows are its own findings.
Open Neural Network Exchange (ONNX) is an open standard for machine learning interoperability. Prior to 1.22.0, RNNShapeInference_opset1_to_6() processes malformed RNN and LSTM nodes from opsets 1 through 6 and GRU nodes from opsets 3 through 6 without verifying that the first input has rank three. The function accesses dim(0) and dim(1) on an undersized protobuf shape, which can read outside the intended repeated field and terminate an application when onnx.shape_inference.infer_shapes() processes an untrusted model that can pass onnx.checker.check_model(). This issue is fixed in version 1.22.0.
This page as .md · .json · in your agent: get_signal("CVE-2026-115919")
NVD data is provided by NIST and is public domain. Posts quoted belong to their authors.