Initial release of the trigger-evaluator skill:

- Evaluates real OpenClaw trigger rules using the current database state.
- Supports "stale_missions_alert" trigger detection based on public.openclaw_* tables.
- Wraps the production script stale_missions_engine.sh for reliability.
- Includes commands to run evaluations and inspect current trigger rows.
- Does not introduce or invent new trigger logic; strictly follows implemented workspace logic.