Openvino Tflite Output Count Oob Poc is a proof-of-concept repository focused on demonstrating a specific heap out-of-bounds read vulnerability in the TensorFlow Lite frontend of OpenVINO. The repository provides both control and malicious model files designed to reproduce the issue, which occurs during model loading when a TFLite operator declares more outputs than the OpenVINO translator returns. This discrepancy can lead to deterministic process termination at load time. The resource is intended for coordinated vulnerability disclosure processes and for use in defensive regression testing. The evidence advises against executing the malicious model in production environments due to the risk of process termination. The repository is released under the Apache 2.0 license. Its primary audience includes security researchers and developers involved in testing or securing machine learning model loading routines, particularly those working with OpenVINO and TensorFlow Lite integration. The tool is delivered as a set of files hosted on a public repository, and there is no indication of a paid or restricted license.
In the Other AI space, Openvino Tflite Output Count Oob Poc takes a focused approach. It helps security researchers and developers test and validate OpenVINO's handling of TFLite output-count vulnerabilities. Openvino Tflite Output Count Oob Poc is an open-source project aimed at AI security researchers. The project is open source (Apache-2.0). It runs on the command line, and it can be self-hosted.
Behind Openvino Tflite Output Count Oob Poc is character-s, and the product first shipped in 2018. The project is developed in the open on GitHub with 10.5k stars and 861 commits in the last 90 days. Among its 5 catalogued features are security testing, poC model, and openVINO compatibility.
Latest indexed changes and source events
character-s/openvino-tflite-output-count-oob-poc verified by the PulseGate indexer
Other apps tracked under the same category.
