Mister Smith Runtime Eval

Use this skill when the ask is to prove what the current Mister Smith runtime actually does on a live run.

This is a runtime-evaluation skill, not an implementation skill and not a planning-only skill.

When To Use

Trigger this skill when the user wants any of the following:

• a real runtime-backed session or task evaluation
• proof that task, session, and autonomy surfaces agree
• a fresh rerun of an earlier proof note on current main
• artifact capture for runtime logs, payloads, and result surfaces
• confirmation of success, collapse, or failure-visible behavior on the supported live path
• an honest comparison between deterministic checks and real runtime behavior

Do Not Use