一句话：展示约束引导的语法模糊器 FANDANGO 能生成100%合法且针对性强的输入，在 GNSS 模块上暴露潜在弱点。
模糊测试广泛用于发现软件漏洞，但传统模糊器难以为复杂输入格式生成合法且有意义的测试用例。基于语法的模糊器虽能保证语法正确，却常常缺乏细粒度控制以触发特定行为。本文展示 FANDANGO 的灵活性与有效性：该先进的语法模糊器结合约束求解，可生成 100% 合法的输入，同时引导生成过程朝向期望的边界情况。作者以 GNSS 模块为案例，说明如何指定约束来探索模块行为。实验表明，FANDANGO 能生成针对性的测试用例并暴露潜在弱点，从而印证约束引导的语法模糊测试在安全测试与可靠性分析中的实际适用性。
Fuzzing is a widely used technique for uncovering vulnerabilities in software systems, but traditional fuzzers often struggle with generating valid and meaningful test cases for complex input formats. Grammar-based fuzzers address this issue by ensuring syntactic correctness, but they frequently lack fine-grained control over generated inputs to trigger specific behaviors. In this paper, we demonstrate the flexibility and effectiveness of FANDANGO,a state-of-the-art grammar-based fuzzer that incorporates constraint solving to produce 100% valid inputs while also guiding the generation process toward desired edge cases. Using a GNSS (Global Navigation Satellite System) module as a case study, we showcase how FANDANGO enables the specification of constraints to explore the module's behavior. Our experiments highlight FANDANGO's ability to generate targeted test cases that expose potential weaknesses. This study reinforces the practical applicability of constraint-guided grammar fuzzing in security testing and reliability analysis.
@inproceedings{paperbot9610,
title = {Personalized Fuzzing: A Case Study with the FANDANGO Fuzzer on a GNSS Module (Short Paper)},
author = {Stephan Neuhaus and Jose Antonio Zamudio Amaya and Andreas Zeller},
booktitle = {ISSTA},
year = {2025},
doi = {10.1145/3713081.3731722}
}