Web-based apps are becoming increasingly popular as the internet grows. Since attackers have focused more on web-based attacks, the security of these apps has become a key problem. Thus, this research study presents a machine learning-based online application plugin to identify and mitigate the most prevalent web vulnerabilities. The plugin detects Proxy IP/VPN, XSS, SQLi, HTTPS malware, and Prototype Pollution assaults. The research develops a complete solution using machine learning and non-machine learning methods. Web traffic data is used to train machine learning models and rule-based algorithms for non-machine learning detection.