Abstract

ObfusQate is the pioneering tool designed to obfuscate quantum programs, enhancing their security by making them resistant to reverse engineering and analysis. The tool employs two main types of obfuscation: quantum circuit obfuscation and control flow obfuscation. Participants will gain hands-on experience using ObfusQate to obfuscate traditional quantum algorithms. We will also demonstrate how the tool can hide malicious code within quantum programs through quantum obfuscation while maintaining semantic equivalence of the code.

Quantum Circuit Obfuscation: We have developed and implemented three different types of circuit obfuscation techniques for the first version of our tool. This approach involves using quantum gates and circuits to obscure the core functionality of quantum algorithms at the circuit level. By employing methods such as inverse gates and identity matrices, ObfusQate creates complex and non-intuitive quantum circuits, making them difficult to interpret and analyze.

Control Flow Obfuscation: We have developed and implemented four different types of code flow obfuscation techniques that will obscure the control flow of the quantum programs. This approach targets the logical sequence of operations within quantum programs. By manipulating the control flow and using quantum-based opaque predicates, ObfusQate obscures the execution path, making it more challenging for adversaries to predict and understand the program’s behavior. Control flow obfuscations are created by utilizing quantum properties like superposition and entanglement, thereby making automatic deobfuscation possible only if quantum principles are broken.

Like other obfuscators from the classical computing world, ObfusQate can be used for both defensive and offensive purposes. It can protect intellectual property by obfuscating quantum programs, preventing adversaries from reverse-engineering proprietary algorithms. Additionally, the tool can be used to hide malicious content within normal quantum programs, potentially bypassing security detections. We will demonstrate the tool’s capability to obfuscate quantum programs without altering their semantics and show how obfuscated malicious code can evade detection by various Large Language Models.