//! miranda-supervisor — Work Order 2's Nemotron-Flash supervisory routing:
//! turn-taking evaluation and dispatching downstream signals to TTS/motion
//! nodes. Also home for the Work Order 1 "Node Warden" concept (localized
//! llama.cpp micro-LLMs per node — see `llamacpp-huggingface-expert` skill).

pub mod turn_machine;