authoring-data-model — SKILL.md

Variant: standard · When to use: authoring a data model document. This skill supplies the producer METHOD and the integrity/queryability BAR; the section STRUCTURE comes from a separate data-model template tool you compose with. Do not restate that template's outline here.

Overview

A data model document is the persistence/domain model of a system's stored data — the design an engineer implements the schema from and everyone who queries the data reads. This skill guides a producer to author one: deriving entities from the upstream feature-spec, detecting the storage paradigm and modeling in it, making integrity rules explicit, and justifying every index by an access pattern. It teaches the how-to — the research method and the quality bar — and composes with a separate template tool that supplies the section structure. The bar is "an engineer can create the schema and write correct queries from it, and the model's integrity rules + tradeoffs are explicit."

It is paradigm-aware: the same modeling rigor (entities, relationships, integrity, justified indexes, stated tradeoffs) applies to relational/SQL, document/NoSQL, graph, wide-column, and key-value stores; only the artifact adapts. Relational + document are the two detailed tracks below; the others are modeled in their own idioms (Step 2).

When to activate

• Authoring or substantially revising a data model / schema design document for a system or feature, given an upstream feature-spec (and an architecture-doc naming the stores, where present).
• The producer needs the method (how to derive entities, model relationships, justify indexes, state tradeoffs) and the bar a buildable, queryable model must meet — not a blank page.

Do NOT activate when: