EvilTwin-Guardian: A Passive Wi-Fi Management-Frame Dataset and AGIF Detection Pipeline for Edge-Based Evil Twin Detection

This dataset accompanies the paper "EvilTwin-Guardian: An AI-Gated Isolation Forest Architecture for Edge-Based Evil Twin Wi-Fi Attack Detection on Raspberry Pi." It contains 351,737 passive Wi-Fi management-frame scan observations collected from a live, multi-band, edge-deployed Raspberry Pi 4 sensor without randomized MAC filtering. The dataset includes the full detection pipeline implementing the Symmetry-Breaking Wireless Feature Taxonomy (SBWFT) across 29 engineered features spanning baseline, temporal, physical, contextual, and security domains.