What problem does it solve? Bootstrapping documentation for an existing data project is tedious: every table needs a permanent spec file, but writing them by hand from scattered SQL models, YAML definitions, and Python schemas takes hours. This Skill scans your existing implementation artifacts and generates a starter spec.md for every table it finds. ## Core Features & Use Cases - Multi-format table discovery: Extracts table definitions from model.yaml files (via the modscape CLI), SQL files (CREATE TABLE and dbt models), and Python files (SQLAlchemy, PySpark, pandas). - Standardized spec output: Writes .modscape/specs/<table-id>/spec.md with Overview, Business Context, Business Rules, and Changelog sections, skipping files that already exist. - Optional model.yaml sync: Can add newly discovered tables to model.yaml and validate the result with modscape validate. - Use Case: You inherit a dbt project with 40 SQL models and no documentation. Run the Skill against models/**/*.sql to bootstrap a spec file per table, then enrich them later through the regular SDD flow. ## Quick Start Ask the AI to run modscape-spec-generate against your model files, for example: generate table specs from model.yaml and all SQL files under models/staging.
