PhishGuard

a Python-based machine learning system designed to detect phishing attacks by analyzing URLs, webpage structure, and embedded content. Built with a decision tree classifier and trained on large datasets of phishing and legitimate links, the model outputs a probability score indicating how likely a URL is safe or malicious. Using Python’s ML and data-processing libraries, I engineered feature-extraction methods such as URL pattern recognition, HTML structure analysis, and keyword detection to strengthen accuracy and reduce evasive threats. The Chrome extension that automatically detects and analyzes the website a user is currently visiting, sending the URL to the backend for instant classification. In addition, we built an Outlook email-scanning feature that inspects incoming messages and flags suspicious links directly in the inbox. These additions extend PhishGuard beyond a standalone tool and into a seamless, real-time protection system across browsers, emails, and local applications.