Abstract:
Phishing attacks have emerged as a significant cybersecurity threat in the digital era, with attackers leveraging fraudulent emails to exploit individuals and organizations. Addressing this challenge, InboxGuard has been developed as a browser extension designed to detect and prevent phishing attempts by analyzing email content and identifying malicious indicators. The project aims to provide a versatile solution that is compatible with mainstream email service providers (ESPs) such as Gmail, Yahoo, and Outlook, ensuring broad applicability across diverse user environments. The extension dynamically extracts email components, including the subject, sender, recipient, and embedded links, and classifies the content into textual data and links for targeted analysis. Embedded links are examined using Google’s Safe Browsing API to identify potential threats, while text-based analysis leverages both the InboxGuard model which is trained by us and the Groq LLM API to evaluate sentiment, tone, and intent based on whether the it is phishing or not. A structured weightage and probability system further enhances the detection process by factoring in key indicators such as recipient count and sender legitimacy, contributing to improved accuracy in identifying phishing emails. Designed for real-time operation, InboxGuard processes emails within a maximum response time of one and a half seconds, regardless of the complexity of the content, ensuring practical and efficient use. Additionally, the extension incorporates an intuitive interface that alerts users to potential threats and provides actionable insights, enabling informed decisions about email interactions. Future enhancements include the development of a comprehensive dashboard to monitor phishing attempts, maintain logs, generate reports, and implement a dedicated database for optimized detection and performance, further solidifying InboxGuard as a reliable solution in the fight against phishing.
