/magic:spec

Manage the workspace data spec for the current data project.

Read the magic-data-lifecycle SKILL.md for the full data spec format and lifecycle context.

Actions (based on arguments or current state):

• No arguments + no spec exists: Infer a data spec from discovery results. Run discovery scripts if needed, draft specs/data-spec.md, present to user for approval.
• No arguments + spec exists: Show current spec status — processing tasks, success criteria, completion progress.
• "evaluate" or "check": Run validation against the spec's success criteria. Produce a compliance report.
• "refine" or "update": Update the spec based on new findings, user feedback, or changed requirements. Record changes in Refinement History.

This triggers Tier 2+ lifecycle. Creating or managing a data spec implies full lifecycle tracking.