define-data-model

Data Model (2-pass)

Shared files: a path written @rules/…, @skills/…, @templates/… or @docs/… is relative to the plugin root, ${CLAUDE_PLUGIN_ROOT} — not to the project being worked on. Read it from there.

Desired Outcome

Produce one deliverable:

1. Data model — reports/02_spec/data-model.md (ENT- IDs):
   • Entities, attributes, and relationships
   • A Mermaid ER diagram of the final model (explicit + implicit)
   • For each Pass-2 addition, the functional-relationship rationale

Invocation

/product:define-data-model [--auto] [--lang=ja|en]

Argument/Flag	Required	Description
--auto	Optional	Derive without elicitation; record ambiguous concepts as TBD
--lang	Optional	Override output language

Decision Criteria

• Entity criteria: a concept becomes an entity only if it is persisted, has identity/key, and receives multiple operations. Prefer value objects; avoid over-normalization.
• Two passes are mandatory. Pass 1 = explicit (UI + features); Pass 2 = implicit via the CRUD matrix — every implicit addition records why (the functional relationship).
• Model the domain, not the schema. Let storage follow the model.
• Stop condition: all explicit concepts captured (Pass 1), the CRUD matrix run to surface implicit entities (Pass 2), and a valid Mermaid ER diagram produced.

Prerequisites

Input	Required/Recommended	Source	If missing/empty
reports/02_spec/feature-list.md	Required	/product:define-features	block with a message — the CRUD matrix needs features
reports/02_spec/ui-mocks/	Required	/product:generate-ui-mock	block with a message — Pass 1 needs screens/forms

Process

1. Read context — features, mocks, work/traceability.json.
2. Pass 1 (explicit) — screen/concept → entity; form fields → attributes; action results → state/events; Command-receiving/Event-emitting aggregate → entity. Apply @rules/product/ui-to-domain.md and @rules/product/ddd-strategic.md.
3. Pass 2 (implicit) — build the feature×entity CRUD matrix; detect many-to-many/history/audit/ state-machine concepts; add them with rationale.
4. Relate — define relationships and cardinalities; keep aggregates small (reference by id).
5. Diagram — render the Mermaid ER diagram of the final model.
6. Append traceability — add ENT- nodes to work/traceability.json with Upstream FEAT-/JOB- references.
7. Record — write the file; append decisions to work/context.md; ask remaining unknowns and log only what stays open (@rules/open-questions.md).

Output

Every Markdown document here opens with the ## ELI5 section of @rules/eli5-summary.md, as options.eli5 (default technical) directs.

reports/02_spec/data-model.md, with an ENT- ID table (Upstream column), Pass-2 rationale notes, and a Mermaid ER diagram.

Reference Materials

Resource	Purpose
@rules/product/ui-to-domain.md	Two-pass derivation, CRUD matrix, entity criteria
@rules/product/ddd-strategic.md	Entity vs value object, aggregates, ubiquitous language

Related Skills

Skill	Relationship
/product:define-features	Upstream — features drive the CRUD matrix
/product:generate-ui-mock	Upstream — screens/forms drive Pass 1
/product:map-domains	Downstream — groups entities into bounded contexts
/product:design-api	Downstream — API resources map to entities
/product:adapt-change	Re-runs this skill when features change