Vigilant Eye

Inspiration We watched a family member nearly fall for a sophisticated phishing email that bypassed all traditional filters. It hit us: current tools build walls, but scams walk through the front door by tricking us. We wanted to build a guide, not just a gate—a tool that makes the invisible tactics of social engineering visible in real-time.

What it does Vigilant Eye is a browser extension that acts as a real-time co-pilot against scams. It actively scans pages as you browse, highlights manipulative language (like fake urgency or too-good-to-be-true offers), warns you about suspicious forms asking for personal data, and explains the risks in simple terms. It turns every potential dangerous moment into a learning opportunity.

How we built it We built a Chromium extension using Manifest V3. The core is a local JavaScript rule engine that scans page text against a curated scamRules.json database. We used Chrome's APIs for scripting and storage. The frontend popup dashboard was built with HTML/CSS, and we designed a non-intrusive alert system that injects warnings directly into viewed pages. The entire prototype runs client-side for privacy and speed.

Challenges we ran into Balancing Alert Precision: Avoiding both false positives (annoying users) and false negatives (missing scams). Tuning our keyword scoring system was a constant challenge.

Performance vs. Thoroughness: Ensuring our page scanner was comprehensive without slowing down the browsing experience required efficient DOM traversal techniques.

The UX of Warnings: Designing alerts that were urgent enough to be noticed but not so alarming they'd be immediately dismissed was a subtle design hurdle.

Accomplishments that we're proud of Creating a fully functional, installable prototype that works in real-time on live websites.

Pioneering an educational-first approach to scam prevention, moving beyond simple blocklists.

Achieving a clean, intuitive user experience that makes complex threat intelligence understandable for anyone.

What we learned Technical: The intricacies of the modern browser extension architecture and the importance of efficient DOM manipulation.

Human-Centered: Effective cybersecurity isn't just about algorithms; it's about understanding psychology and designing interventions that respect the user's mindset and workflow.

Teamwork: How to divide a complex project (rule engine, content injection, UI dashboard) into parallel tasks that integrated seamlessly under time pressure.

What's next for Vigilant Eye Crowdsourced Intelligence: Implementing a "Report Scam" button to build a community-driven threat database.

Advanced Detection: Integrating a lightweight machine learning model to detect novel scam patterns based on linguistic style.

Organizational Tools: Building a dashboard for schools or small businesses to track threats and conduct training simulations for their users.