Systematic Security and Privacy Auditing of the Android Ecosystem via Automated System-Level Re-Hosting
Abstract
The Android ecosystem's highly fragmented, multi-tiered supply chain introduces significant security and privacy risks, often resulting in custom vendor firmware containing silent vulnerabilities, over-privileged pre-installed applications, and undocumented data collection behaviours. Traditionally, researchers rely on two primary methodologies to audit these devices: static and dynamic analysis. While static analysis scales well, it suffers from high false-positive rates and fails to capture complex runtime interactions. Conversely, dynamic analysis is severely bottlenecked by the lack of physical hardware scale and the limitations of stock emulators in executing heavily customized vendor components. To bridge this gap, this dissertation proposes a systematic, automated framework for Android supply-chain security and privacy auditing. The core of our approach lies in automated system-level re-hosting architectures, allowing proprietary, vendor-customized system images to be executed and analysed dynamically within instrumented, virtualized environments. Through this unified methodology, we aim to scale dynamic vulnerability discovery, trace complex data-leakage paths in pre-installed software, and provide the broader mobile security community with an open, reproducible framework for supply-chain auditing.