Skip to content

Latest commit

 

History

History
103 lines (65 loc) · 3.9 KB

README.md

File metadata and controls

103 lines (65 loc) · 3.9 KB
Raw

DeepSQLi

DeepSQLi is a project designed to detect SQL injection (SQLi) attacks using deep learning models. It provides a Prediction API, a Flask application that tokenizes and sequences SQL queries, and then predicts whether a query is an SQL injection.

Project Overview

The Prediction API handles both tokenization and prediction in a single endpoint, simplifying the detection workflow. It processes incoming SQL queries and determines if they contain SQL injection patterns.

GatewayD Integration

This project can be integrated with GatewayD using the GatewayD SQL IDS/IPS plugin. The plugin acts as a middleware between clients and the database, intercepting SQL queries and sending them to the Prediction API for analysis. If a query is classified as malicious by the Prediction API, the plugin blocks the query; otherwise, it forwards the query to the database.

Architecture

flowchart TD
    Client <-- PostgreSQL wire protocol:15432 --> GatewayD
    GatewayD <--> P["Prediction API"]
    P --loads--> SM["SQLi Model"]
    P --threshold: 80% --> D{Malicious query?}
    D -->|No: send to| Database
    D -->|Yes: terminate request| GatewayD
Loading

Models and Tokenization

Model Versions

  • LSTM Models: The first two models are LSTM-based and are trained using different datasets:

    • sqli_model/1: Trained on dataset v1 (for historical purposes, will remove in the future).
    • sqli_model/2: Trained on dataset v2 (for historical purposes, will remove in the future).

  • CNN-LSTM Model: The third model, sqli_model/3, is a hybrid CNN-LSTM model with a custom SQL tokenizer to improve performance on SQL injection patterns (recommended).

Tokenization

The Prediction API performs tokenization internally:

  • Default Tokenizer (Models 1 and 2): Uses Keras’s Tokenizer for general tokenization.
  • Custom SQL Tokenizer (Model 3): A custom tokenizer designed to handle SQL syntax and injection-specific patterns.

Installation

Docker Compose

To start the Prediction API with Docker Compose:

docker compose up --build -d

To stop and remove the containers:

docker compose down

Docker (Manual Setup)

Build Images

docker build --no-cache --tag prediction-api:latest -f Dockerfile .

Run the Container

docker run --rm --name prediction-api -p 8000:8000 -d prediction-api:latest

Usage

Once the Prediction API is running, use the predict endpoint to classify SQL queries.

Prediction API

curl 'http://localhost:8000/predict' -X POST -H 'Content-Type: application/json' \
--data-raw '{"query":"SELECT * FROM users WHERE id=1 OR 1=1;"}'

Response Format

The response includes the prediction (1 for SQL injection, 0 for legitimate query) and confidence score. Note that the confidence score is only available for the CNN-LSTM model using the Prediction API.

{
    "confidence": 0.9722,
}

Contributing

We welcome contributions from everyone. Just open an issue or send us a pull request.

License

This plugin is licensed under the Affero General Public License v3.0.

Acknowledgements

This project is inspired by the SQL Injection Dataset project by Abu Syeed Sajid Ahmed and the A CNN-BiLSTM based Approach for Detection of SQL Injection Attacks conference paper by Neel Gandhi, et al.