What problem does it solve? Applying seed data to a database is risky: non-idempotent scripts fail on re-runs, destructive statements can wipe data, and there is often no rollback plan. This Skill enforces a safe, repeatable seeding workflow with pre-flight checks, snapshots, and verification. ## Core Features & Use Cases - Idempotency Validation: Scans the seed file for dangerous patterns like TRUNCATE or DELETE and confirms ON CONFLICT usage before execution. - Snapshot and Rollback: Creates a timestamped pg_dump snapshot before seeding so you can restore if the seed fails. - Verification and Logging: Runs row-count verification queries and appends an entry to a SEED_LOG.md audit file after each run. - Use Case: You have a new categories seed file for your Supabase project. Run the skill to validate it is idempotent, snapshot the current schema, apply the seed in a transaction, and log the operation. ## Quick Start Ask the agent to apply the seed file at supabase/seeds/categories.sql to the database using the db-seed workflow.