# LOC-002-LAB1 — Tauri Command Boundary & Mediation Contract v1

This is an offline synthetic lab. Implement `evaluate_case` in `starter.py` and use the supplied fixtures to build a bounded command-mediation matrix.

Keep these layers separate:

1. Tauri runtime-authority exposure (origin/capability/permission),
2. explicit command registration and typed input validation,
3. MAKNOW domain authorization and bounded resource scope,
4. idempotency/audit behavior for state-changing operations.

Run `python run_checks.py` for learner-visible structural checks. The evaluation oracle is intentionally not in the learner root. Do not use real credentials, production paths, privileged OS changes, or external targets.