Cybershield AI: A Hybrid Framework for Malicious URL Detection using Heuristic Analysis and a Proposed ML Architecture

The proliferation of malicious Uniform Resource Locators (URLs) represents a significant threat to cybersecurity. While machine learning (ML) models offer high accuracy,traditional heuristic and rule-based systems provide a trans parent, low-latency first line of defense. This paper introduces Cybershield AI, a hybrid framework for real-time URL analysis. We present the design, implementation, and evaluation of the first component of this framework: a high-speed heuristic engine based on lexical and rule-based checks. This engine was validated on a test set of 40 URLs, achieving 87.5Index Terms—Malicious URL Detection, Heuristic Analysis,Machine Learning, Cybersecurity, Threat Analysis, ExplainableAI (XAI)