| name | data-model |
| description | Use when creating or revising Lindy data models: durable, simple schemas for JSON files, JSON schemas, API payloads, database schemas, SQL tables, migrations, or any structured set of keys, fields, or columns. Acts as a lightweight wrapper around Schemator and its bundled schemator skill. |
Data Model
Use this skill when the goal is to create a Lindy data model: a schema that is
small, durable, boring in the best way, and likely to survive changing product
details because it captures stable concepts instead of temporary assumptions.
This skill is only a wrapper. Schemator is the source of truth for reviewing
schemas, simplifying fields, checking naming, and producing auditable reports.
How To Use It
- Use Schemator and its bundled
schemator skill for the actual workflow. If
the skill is installed, invoke $schemator.
- Start from a real draft schema or proposal. If no draft exists yet, make the
smallest useful draft first, then hand it to Schemator.
- Provide project/task context before review so Schemator can judge names,
field purpose, constraints, and borrowed vocabulary in context.
- Run Schemator for extraction, review, simplification, convergence, diffing,
or reporting as appropriate for the task.
- Treat Schemator output as a candidate result and do the final human/product
pass required by the
schemator skill.
Schemator Reference
Install Schemator from npm:
npm install -g @osolmaz/schemator
Use the bundled Schemator skill as the detailed instruction source:
schemator --skill show schemator
schemator --skill export schemator | npx skillflag install --agent codex
Do not duplicate Schemator's review rules, report checklist, prompt guidance, or
schema criteria in this skill. Update Schemator and its bundled skills instead.