loop-audit

Audit a command/skill library (or a single workflow) against the Loop Engineering five-moves rubric — discovery, handoff, verification, persistence, scheduling — plus the four silent costs (verification debt, comprehension rot, cognitive surrender, token blowout) and the generator/evaluator separation test. Use whenever the user wants to know which of their commands are real self-running loops vs. human-in-the-loop scripts, or asks to "loop audit", "audit my loops", "audit my commands/skills", "score my workflows", "are these actually loops", "which commands are real loops", "five moves audit", "loop maturity", "loop engineering review", "where are my automation gaps", or hands over a slash-command/skill directory and wants to know what to automate or harden next. Prefer this over repo-doctor or app-audit when the unit of analysis is loops/commands/agentic workflows rather than application code or runtime behavior.