CyberSentinel – AI-Powered Phishing Detection System

CyberSentinel is a full-stack cybersecurity web application that detects malicious URLs using a hybrid approach combining:

• Machine Learning (Random Forest)
• VirusTotal API
• Content Analysis
• Community Feedback + Retraining

It also includes a Chrome Extension for real-time protection.

---

Features

• Real-time URL phishing detection
• Hybrid analysis (ML + API + Content)
• Admin panel for feedback review
• Model retraining system
• Chrome extension integration
• JWT-based authentication

---

Tech Stack

Frontend

• React (Vite)
• Tailwind CSS

Backend

• Flask (Python)
• Flask-JWT-Extended

Machine Learning

• scikit-learn
• pandas, numpy

Database

• SQLite / PostgreSQL

---

Installation Guide

1. Clone the Repository

git clone https://github.com/yourusername/CyberSentinel.git
cd CyberSentinel
cd backend

Windows

python -m venv venv
venv\Scripts\activate

Mac/Linux

python3 -m venv venv
source venv/bin/activate

Dependencies:

pip install -r requirements.txt

Download Dataset & Train Model:

python ml/download_dataset.py
python ml/train_model.py

Add Environment Variables

Create a .env file inside the backend folder:

VIRUSTOTAL_API_KEY=your_api_key_here
SECRET_KEY=your_secret_key

Get VirusTotal API key from: https://www.virustotal.com/gui/join-us

Start Backend Server

python app.py

Backend will run on: http://localhost:5000

Frontend Setup

Open New Terminal & Navigate to Root

cd ..

Install Dependencies

npm install

Start Frontend

npm run dev

Frontend will run on: http://localhost:5173

Chrome Extension Setup

Open Chrome and go to: chrome://extensions/ Enable Developer Mode Click Load Unpacked Select the extension folder from this project

Default Admin Login

Username: admin Password: admin123

How it works

1. User enters a URL or visits a website
2. Extension / frontend sends URL to backend
3. Backend performs:
   • ML prediction
   • VirusTotal check
   • Content analysis
4. Results are combined into a threat score
5. Final verdict is returned (Benign / Suspicious / Phishing)

Model Retraining

• Users submit feedback
• Admin approves feedback
• Approved data is added to dataset
• Model is retrained using:
• POST /api/admin/retrain-model

Authors

Connect with Us:

Mohataseem Khan LinkedIn • GitHub

Rehan Khan LinkedIn • GitHub

Saad Shaikh LinkedIn • GitHub

Ansari Hussain LinkedIn • GitHub