What problem does it solve? Deciding which fields an entity should have and where each field should live (table column, config blob, or runtime-computed) is error-prone, leading to bloated schemas, redundant fields, and painful migrations. This Skill turns a PRD plus an optional codebase into a structured audit that answers what should exist, what exists now, and what must change. ## Core Features & Use Cases - Five-Cut Field Classification: Every field passes through existence, store-vs-compute, column-vs-config, system-vs-content, and stored-form decisions, each framed as a cost-to-change judgment. - 8-Column Audit Table: Produces a target-vs-current comparison per entity covering type, classification, field nature, existence justification, current source, and required change. - Migration Change List: Groups renames, new fields, removals, moves, and API gaps into an actionable plan, plus design-decision justifications. - Use Case: Paste a PRD and a Prisma or SQLAlchemy model and ask whether the fields are right; receive a full audit report showing which columns to keep, which to move into a jsonb config blob, and which to compute at runtime. ## Quick Start Paste your PRD and current database model, then ask the agent to audit the entity fields and produce the data model audit report.
