This project is a proof-of-concept (POC) and is not actively maintained. It is intended for research and educational purposes only.

If you discover a security vulnerability in this project, please follow these steps:

1. Do not open a public GitHub Issue with exploit details.
2. Instead, email the maintainer directly or open a GitHub Issue with a general description only (no exploit code or payload).
3. Allow up to 30 days for a response before any public disclosure.

This tool is designed to detect SQL Injection and XSS attacks — it is not designed to be deployed as a production-grade firewall. Known limitations include:

• The neural network is trained on a fixed dataset and may not generalise to novel attack patterns.
• No input sanitisation is applied to the HTTP traffic loader itself.
• The saved model file (/tmp/ann.db) has no integrity verification and should not be used in untrusted environments.

This project is provided as-is under the MIT License. The authors accept no liability for misuse or deployment in production systems.