Poster: Detecting Intent Inconsistency in Mobile Apps:A Focus on Interaction Flow Contamination

Abstract

The commercialization of the mobile Internet has driven many apps to adopt deceptive interaction designs (referred to as interaction flow contamination), which misalign system behaviors with user true intent and may harm user interests. Existing detection approaches suffer from semantic comprehension gaps, insufficient behavioral traceability and lack of quantitative evaluation frameworks. In this paper, we study 200 mainstream Android apps crawled from public app markets across 8 categories, and propose an intent-consistency detection framework. We formalize intent consistency as a multi-dimensional model covering UI semantics, sensor states and page transitions. Since deceptive behaviors can exist in both static UI layouts and runtime dynamic hooks, we build a dual-track prototype combining ADB-based static scanning and Frida dynamic instrumentation (spawn-mode injection for early pre-start monitoring). Our evaluation shows that 69% of tested apps exhibit interaction-flow contamination. The prototype achieves 94.9% precision; static scanning latency stays under 300 ms, with an average per-app detection time of 7.6 seconds. The results demonstrate that intent consistency can serve as a quantitative compliance-auditing metric for mobile interaction behaviors.