Instructions file GitHub Copilot
Guidelines for building safe, governed AI agent systems. Apply when writing code that uses agent frameworks, tool-calling LLMs, or multi-agent orchestration to ensure proper safety boundaries, policy enforcement, and auditability.