What problem does it solve? When adopting Drizzle ORM on a project that already has a live database, you need a schema.ts that matches reality without hand-writing it. This Skill reverse-engineers the existing database into a Drizzle schema, relations file, and snapshot, and decodes the result manifest or connection failures. ## Core Features & Use Cases - Database Introspection: Connect to a live PostgreSQL, MySQL, SQLite, Turso, CockroachDB, SingleStore, or MSSQL database and emit schema.ts, relations.ts, and a snapshot under the configured out directory. - CLI and SDK Forms: Run drizzle-kit pull with flags like --casing, --breakpoints, and --output json, or call the pull() SDK function which returns the same ok/error envelope. - Manifest Decoding: Interpret the paths manifest (schemaPath, snapshotPath, relationsPath, migrationPath) and the credential-redacted database_driver_error envelope. - Use Case: You inherit a legacy PostgreSQL database and want to manage it with Drizzle. Run drizzle-kit pull --output json to generate drizzle/schema.ts and drizzle/relations.ts, then optionally use --init to write the initial migration. ## Quick Start Ask the AI to introspect your existing database with drizzle-kit pull and generate the Drizzle schema files into your configured out directory.