README.md

metadata

title: AI Code Security Scanner
emoji: 🛡️
colorFrom: blue
colorTo: indigo
sdk: streamlit
sdk_version: 1.52.2
app_file: app.py
pinned: false
app_port: 8501
short_description: AI Python security scanner using CodeBERT and rules.

AI Code Security Scanner: Real-time Vulnerability Detection

An advanced security orchestration tool that detects critical vulnerabilities (SQLi, XSS, Command Injection) using a Hybrid Analysis Engine: combining static rule-based heuristics with deep learning CodeBERT models.

Live Demo: AI Security Scanner on Hugging Face

---

Core Engineering Features

• Hybrid Detection Engine: Utilizes AST (Abstract Syntax Tree) parsing for rule-based detection and fine-tuned CodeBERT (Deep Learning) for semantic vulnerability pattern matching.
• Automated Remediation: Generates context-aware "Fix Suggestions" using AI to provide secure code alternatives for identified risks.
• Security Dashboard: Real-time visualization of security scores and vulnerability distribution across large codebases.
• Batch Processing: Optimized for high-throughput scanning of multiple Python files simultaneously using Python's multiprocessing capabilities.
• REST API Architecture: Built with a FastAPI backend, allowing for easy integration into existing CI/CD pipelines as a security gate.

Technical Stack

• Deep Learning: CodeBERT (Transformers), PyTorch
• Analysis Engine: Python AST, Custom Rule-based Heuristics
• Backend/Frontend: FastAPI, Streamlit
• Deployment: Docker, Hugging Face Spaces (Debian Linux)

Installation

1. Clone the repository:

   git clone [https://github.com/mubi0613/AI-Code-Security-Scanner.git](https://github.com/mubi0613/AI-Code-Security-Scanner.git)
   cd AI-Code-Security-Scanner
   

2. Setup environment:

   python -m venv venv
   
   # On Windows:
   .\venv\Scripts\activate
   # On Mac/Linux:
   source venv/bin/activate
   

3. Install dependencies:

   pip install -r requirements.txt
   

4. Run the application:

   streamlit run app.py`
   

Docker Deployment

• Build and run with Docker Compose
• docker-compose up --build

Access at:

• Web UI: http://localhost:8501
• API Docs: http://localhost:8000/docs

Project Structure

AI-Code-Security-Scanner/

• ├── app.py # Main Streamlit Dashboard & UI Logic
• ├── api_backend.py # FastAPI Server for REST API Access
• ├── combined_detector.py # Hybrid Logic (Rules + CodeBERT Model)
• ├── rule_detector.py # Static Analysis & AST-based Heuristics
• ├── fix_generator.py # AI-powered Remediation Engine
• ├── saved_model/ # Fine-tuned CodeBERT Weights & Config
• ├── docker-compose.yml # Container Orchestration for UI & API
• ├── requirements.txt # Project Dependencies
• └── README.md # Technical Documentation

Architecture Diagram

graph LR
    subgraph "Frontend Layer"
        A[Streamlit UI] --> B[User Code Input]
    end

    subgraph "Analysis Engine"
        B --> C{Hybrid Scanner}
        C --> D[Static Rule Engine]
        C --> E[CodeBERT DL Model]
    end

    subgraph "Remediation & Reporting"
        D & E --> F[Vulnerability Report]
        F --> G[AI Fix Generator]
        G --> H[Final Security Report]
    end