README.md

metadata

title: SecurityChatbot
emoji: 📚
colorFrom: gray
colorTo: pink
sdk: docker
pinned: false

Security Awareness Chatbot 🔐🤖

A modular, AI-powered backend that simulates social engineering attacks and provides interactive quizzes to train users on cybersecurity best practices—all in a friendly Egyptian Arabic persona!

🌟 What Does This Bot Offer?

1. 🛡️ Interactive Security Assistant Talk to a cybersecurity expert AI. Ask any question about digital safety, and get answers in simple, casual Egyptian Arabic.

2. 🕵️‍♂️ Social Engineering Simulation (Roleplay) Users can type /train to enter a simulated attack scenario.

   • Scenarios: Bank employee, Delivery Courier, IT Support, WhatsApp Support, Lottery.
   • Difficulty Levels: Easy, Medium, Hard (ranging from obvious to highly manipulative).
   • Mechanism: The AI plays the attacker, trying to extract sensitive info. A separate AI (Evaluator) acts as a referee to catch if the user leaks data or successfully defends themselves for 4 turns.

3. 📝 Dynamic Cybersecurity Quizzes Users can choose the Quiz mode to get a dynamically generated Multiple Choice Question (MCQ) about cybersecurity topics. The bot evaluates their answer and provides a friendly explanation.

4. ⚡ High Reliability (Fallback Logic) Powered by OpenAI (GPT-4o-mini) and Groq (Llama-3), with built-in automatic fallback to Google Gemini (1.5-flash) if the primary models experience downtime.

---

📁 Project Structure

securityChatbot/
├── app/
│   ├── __init__.py
│   ├── main.py
│   ├── core/
│   │   ├── __init__.py
│   │   ├── config.py
│   │   └── prompts.py
│   ├── models/
│   │   ├── __init__.py
│   │   └── schemas.py
│   ├── routers/
│   │   ├── __init__.py
│   │   └── webhook.py
│   └── services/
│       ├── __init__.py
│       ├── llm_service.py
│       └── state_manager.py
├── Dockerfile
├── .dockerignore
├── pyproject.toml
└── README.md

Setup & Run Locally

uv sync
uv run uvicorn app.main:app --reload --host 0.0.0.0 --port 8000

API Endpoints

Method	Path	Description
GET	/	Health check
POST	/webhook	Main chat endpoint

Webhook Payload

{
  "user_id": "user123",
  "message": "Hello"
}

Commands

• /train — Start the social engineering simulation