Contributing

Contributions are welcome, especially:

• Clarifying requirements and non-goals
• Tightening/validating the schema for LLM reliability (low hallucination risk)
• Adding interaction scenarios and edge cases (staleness, conflicts, low-confidence resolution)
• Improving implementer guidance and examples across channels (chat/email/UI)

If you’re proposing a change, please include:

• The motivation/use case
• The fields/behavior impacted
• How the change affects auditability, privacy/visibility, and multi-party resolution

Open a discussion or pull request on GitHub.