libannabil/phishguard-kenya

# PhishGuard Kenya 🛡️ A real-time phishing URL detection and reporting platform built for the Kenyan digital ecosystem. PhishGuard combines machine learning with a hard rule engine to detect phishing sites targeting M-Pesa, KRA, banking portals, and other commonly impersonated Kenyan and global brands. --- ## Features - **URL Scanner** — Paste any URL and get an instant verdict: Safe, Suspicious, or Phishing - **ML + Rule Engine** — Random Forest classifier combined with 11 deterministic security rules (catches typosquatting, brand impersonation, homoglyph attacks, and more) - **Fuzzy Brand Matching** — Detects typosquatted domains like `microsft.com`, `safaric0m.co.ke`, `paypa1.com` - **Community Reporting** — Logged-in users can report phishing URLs for admin review - **Threat Database** — Public dashboard of confirmed phishing threats - **Admin Panel** — Approve/reject community reports, manually add threats - **Education Centre** — Guides and interactive quiz on spotting phishing attacks - **JWT Authentication** — Secure login/register with bcrypt password hashing --- ## Tech Stack | Layer | Technology | |-------|-----------| | Frontend | Vanilla HTML/CSS/JS (single file) | | Backend | Python 3, Flask, Flask-JWT-Extended | | Database | MySQL (SQLAlchemy ORM) | | ML Model | scikit-learn Random Forest | | Auth | Flask-Bcrypt + JWT | --- ## Project Structure ``` phishguard/ ├── backend/ │ ├── routes/ │ │ ├── analyze.py # Core URL analysis endpoint + rule engine │ │ ├── auth.py # Register, login, /me │ │ ├── reports.py # Community phishing reports │ │ ├── threats.py # Public threat database │ │ └── admin.py # Admin moderation endpoints │ ├── models.py # SQLAlchemy database models │ ├── extensions.py # Flask extensions (db, bcrypt, jwt) │ └── app.py # App factory ├── ml/ │ ├── features.py # 30-feature URL extractor │ ├── train_model.py # Model training pi …