Anti-Phish Guard: Intelligent Browser Extension for Real-Time Phishing Detection and Prevention
Conventional web browsing places users at signify-can’t risk of phishing attacks, in which bad actors create malicious websites to deceptively render them vulnerable to the capture of sensitive information, typically through fake links and interfaces. Manual verification techniques are often slow, unreliable, and do not usually identify advanced phishing attacks. To address these issues, this paper describes the design and use of Anti-Phish Guard (APG), an intelligent web browser extension that provides real-time detection and protection against phishing attacks. The system is implemented in modern web technologies, including HTML, CSS, JavaScript, Manifest V3 format, and Node.js based backend APIs. APG automatically scans URLs during page loading and during user interaction using heuristic checks, domain reputation analysis, and pattern-based detection. Untrusted links get flagged, and users receive alerts in real-time through the browser notifi-cation interface. The architecture uses the IP address metadata to find link authenticity and detects potential attackers, while also logging their patterns of attack. The cloud-based backend built using Node.js and MongoDB detects and dedicates known blacklisted URLs, stores records of attack logs, and syncs recent detections. Experimental evaluations show that the detection of phishing URLs is highly accurate and there are few false positives in detection patterns, demonstrating seamless use into normal workflows for browsing. The proposed solution offers a zero-trust model for phishing URL detection but at a secure, scalable, and user-friendly protective layer within the context of social media.
Authors
- Goutham R
- Gagandeep S
- Adarsha
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-07-30
- DOI
- https://doi.org/10.5281/zenodo.21700942
- Primary Topic
- Spam and Phishing Detection
- Type
- article
- Field-Weighted Citation Impact
- 0.00