docs/ai/multi-agent-control.md

# Multi Agent Control

This document explains multi agent control for maintainers. Operational authority remains in the master prompt and focused .ai rules. Use evidence, bounded work, progressive disclosure, and proof-based termination.