PHISHFINDER: A MULTIPRONGED PHISHING DETECTOR PLUGIN FOR CHROMIUM BROWSERS
Mentor
Anderson Scot
Document Type
Thesis
Publication Date
2025
Abstract
Phishing attacks and malicious URLs continue to pose significant threats to internet users, exploiting both technical and human vulnerabilities. While numerous standalone tools exist to detect and block such threats, each has limitations in detection accuracy and user accessibility. This project presents PhishFinder, a Chromium-based browser extension that integrates multiple malware detection APIs into a unified, user-friendly platform. By combining the strengths of tools such as VirusTotal, Google Safe Browsing, and local WHOIS and SSL certification validation, the extension provides real-time, layered protection against harmful URLs without requiring advanced technical knowledge from users. The system is designed to deliver immediate feedback and block access to flagged sites while preserving user privacy through local processing and anonymization techniques.
A weighted-scoring system combines evidence from each source to classify sites as safe, suspicious, or malicious. Testing used the Brave browser, a chromium based platform, on curated datasets of safe, unsafe, and ambiguous URLs. PhishFinder achieved an overall accuracy of approximately 98.8 percent, correctly classifying 100 percent of safe and unsafe URLs and 95 percent of ambiguous URLs once VirusTotal had incorporated prior submissions. Performance analysis showed a 2.95 second average latency in serial execution and 0.28 seconds in parallel execution representing a 90 percent reduction in response time with identical accuracy.
Fifteen user participants evaluated the system and reported high usability, with over 90 percent describing the extension as intuitive and non-intrusive. These results confirm that a multi-sourced, AI-enhanced browser-level approach can substantially improve phishing detection accuracy and speed while maintaining privacy and accessibility.
Recommended Citation
Illangovan, Daryl and Anderson, Scot, "PHISHFINDER: A MULTIPRONGED PHISHING DETECTOR PLUGIN FOR CHROMIUM BROWSERS" (2025). MS in Computer Science Project Reports. 14.
https://knowledge.e.southern.edu/mscs_reports/14