Skip to content

Contributing to AI Runtime Behaviour Security

Thank you for your interest in contributing to this project. This document provides guidelines for contributions.

How to Contribute

Reporting Issues

If you find an error, omission, or have a suggestion:

  1. Check existing issues — Someone may have already reported it
  2. Open a new issue — Use the appropriate template
  3. Be specific — Include the document, section, and what you think should change
  4. Provide context — Why does this matter? What's the impact?

Submitting Changes

  1. Fork the repository
  2. Create a branch — feature/your-feature-name or fix/issue-number
  3. Make your changes — Follow the style guide below
  4. Test your changes — Ensure links work, markdown renders correctly
  5. Submit a pull request — Reference any related issues

What We're Looking For

Highly valued contributions:

Type Examples
Production experience Real-world implementation lessons, cost data, failure modes
Regulatory insight Regulatory feedback, compliance interpretations, examination findings
Platform expertise Implementation guides for specific platforms
Academic research Empirical validation, formal analysis
Corrections Factual errors, outdated information, broken links

Also welcome:

  • Clarifications and improved explanations
  • Additional examples and use cases
  • Translations (discuss first)
  • Tooling and automation

Style Guide

Writing Style

  • Clear and direct — Avoid jargon where possible
  • Practical — Focus on what practitioners need
  • Honest — Distinguish between proven and theoretical
  • Consistent — Match the tone of existing content

Markdown Conventions

# Document Title (H1 - one per document)

## Major Section (H2)

### Subsection (H3)

**Bold** for emphasis on key terms
*Italic* for introducing new concepts
`Code` for technical terms, control IDs, file names

| Column 1 | Column 2 |
|----------|----------|
| Data | Data |

- Bullet lists for unordered items
1. Numbered lists for sequences

> Blockquotes for important callouts

Control Naming

  • General AI controls: AI.x.x (e.g., AI.7.1 Input Guardrails)
  • Agentic controls: AG.x.x (e.g., AG.2.2 Circuit Breakers)
  • Always include the control name after the ID

File Naming

  • Lowercase with hyphens: my-document-name.md
  • Numbered prefixes for ordered content: 01-introduction.md
  • SVG for diagrams: diagram-name.svg

Review Process

  1. Automated checks — Links, markdown formatting
  2. Maintainer review — Content accuracy, style consistency
  3. Feedback — We may request changes or clarification
  4. Merge — Once approved, your contribution will be merged

Review Criteria

Criterion What We Check
Accuracy Is the content factually correct?
Relevance Does it belong in this framework?
Clarity Is it easy to understand?
Consistency Does it match existing style and structure?
Completeness Is it thorough enough to be useful?

Types of Contributions

Documentation Improvements

  • Fix typos, grammar, unclear explanations
  • Add examples or clarifications
  • Update outdated information
  • Improve cross-references

New Content

Before starting significant new content:

  1. Open an issue to discuss the idea
  2. Get feedback on scope and approach
  3. Coordinate to avoid duplicate effort

Areas where new content is welcome:

  • Platform-specific implementation guides
  • Industry-specific guidance (beyond banking)
  • Regional regulatory mappings
  • Worked examples and case studies

Diagrams and Visualisations

  • SVG format preferred
  • Consistent colour palette (see existing diagrams)
  • Clear, readable at reasonable zoom
  • Include alt text descriptions

Code of Conduct

This project follows a standard code of conduct. In summary:

  • Be respectful — Treat others as you want to be treated
  • Be constructive — Focus on improving the content
  • Be patient — Maintainers are volunteers
  • Be inclusive — Welcome diverse perspectives

See CODE_OF_CONDUCT.md for details.

Recognition

Contributors will be recognised in:

  • Git commit history
  • Contributors list (for significant contributions)
  • Release notes (for major contributions)

Questions?

If you're unsure about anything:

  1. Check the existing documentation
  2. Open an issue with your question
  3. Tag it with question

We're happy to help guide your contribution.

AI Runtime Behaviour Security, 2026 (Jonathan Gill).