AI-DRIVEN FRAMEWORK FOR AUTOMATED ANOMALY AND VULNERABILITY DETECTION IN WEB APPLICATIONS

Modern web applications face challenges like dynamic behaviors and expansive attack surfaces, complicating early bug detection.Traditional scanning and manual testing methods are often inadequate due to their lack of adaptability in complex environments.This paper introduces an AI-driven automated bug detection framework that combines systematic reconnaissance and intelligent vulnerability analysis.It gathers application artifacts such as URLs, request parameters, JavaScript files, and subdomain metadata, which are then normalized to eliminate redundancy.A hybrid AI detection engine utilizes pattern recognition, heuristic rules, and machinelearning techniques to discover anomalous behaviors and identify security vulnerabilities.Detected bugs undergo severity analysis and confidence scoring to prioritize them based on their potential impact and reliability.The framework features a web interface for real-time scanning, visualization of analysis phases, and automated vulnerability reporting.Experimental results indicate that this approach enhances detection accuracy, reduces false positives, and lowers the need for manual intervention compared to traditional methods, ultimately improving software security, reliability, and development efficiency while facilitating proactive vulnerability management.