CONTRIBUTING.md

Contributing to this Repository

Thank you for your interest in making AI safer. We welcome contributions!

The Architecture Philosophy

This is not a standard Python library. It follows a strict "Bicameral Architecture":

1. Layer 1 (Prompt): Must isolate "Reasoning" from "Result".
2. Layer 2 (Sanitizer): Must treat all user input as untrusted data.
3. Layer 3 (Validator): Must deterministically verify the output.

If your Pull Request breaks this separation of concerns, it will be rejected.

Development Guidelines

1. New Features

If you add a new capability, you must update all three layers:

• Update the Prompt (The Beak/Door/Quill) to handle the intent.
• Update the Validator (The Tentacles/Tumblers/Lens) to verify the new output.

2. Red Team Testing (Mandatory)

Every new feature must include a Hostile Test Case.

• Example: If you add support for INSERT statements, you must add a test proving it rejects INSERT INTO system_tables.
• Add your tests to the tests/ directory using pytest.

3. Style

• Follow PEP 8 standards.
• Do not include API keys in your commits.

Getting Started

1. Fork the repo.
2. Install dependencies: pip install -r requirements.txt
3. Run tests: pytest
4. Submit your Pull Request.