Key Features: - AI-powered page scanning with Gemini - DOM, script, network, and download monitoring - Popup verdict: SAFE / SUSPICIOUS / DANGEROUS - User-managed API key (no backend or tracking) - Built with Manifest V3 + prompt engineering - Human-readable explanation for threats
I wanted to explore how browser security could evolve beyond static blacklists by integrating *real-time AI analysis* natively into browsing.
GitHub: https://github.com/pradumon14/trinetra Whitepaper included in the repo Feedback and contributions welcome!
Would love to hear your thoughts, feedback, or improvements.