What problem does it solve? Designing database schemas that enforce data integrity is error-prone: missing foreign keys, wrong data types for money or timestamps, unsafe migrations that lock large tables, and engine-specific quirks (SQLite's disabled foreign keys, MySQL's ignored CHECK constraints before 8.0.16) silently corrupt data or cause outages. ## Core Features & Use Cases - Engine-aware schema review: Resolves the database engine and version from project evidence (drivers, migration tooling, connection strings) before judging any DDL, and gates recommendations on version-specific features. - Severity-ordered design catalog: Covers primary keys, foreign keys, NOT NULL and CHECK constraints, UNIQUE constraints, data types, normalization, naming, and schema-change safety, each traced to the official PostgreSQL, MySQL, or SQLite manual. - Migration safety analysis: Evaluates ALTER TABLE operations per engine (PostgreSQL rewrites, MySQL InnoDB instant DDL, SQLite twelve-step rebuild) so a correct target schema is not reached through a locking migration. - Use Case: Point it at a migrations directory in a Rust/sqlx or Go project and receive a structured review report tagging each finding as integrity, static, requirement, or migration, with manual citations. ## Quick Start Ask the AI to review the SQL migration files in this repository for schema design and integrity issues using the appropriate database engine.