Single Source Of Truth Guard

The problem

Database schemas, TypeScript types, and application constants inevitably drift apart. Even with "no magic strings" rules, developers frequently hardcode string literals in webhooks, admin panels, or cron jobs to save time, creating silent bugs that type checkers cannot catch.

What it does

• Enforces a strict hierarchy where the database schema is the sole authority for data shape.
• Ensures every enum-like value (statuses, roles, tiers) is defined in exactly one constants file.
• Identifies and replaces raw string literals or duplicate inline types with centralized imports.
• Mandates that schema migrations and type updates occur in the same atomic change.
• Audits the codebase for hidden "magic string" violations in edge-case locations like webhook handlers and seed data.

Frameworks & tools

Designed for TypeScript environments using any SQL or NoSQL database. Works with migration tools like Prisma, Drizzle, or TypeORM.

Why this beats prompting it yourself

Generic prompts often allow "good enough" local overrides or inline types that pass the compiler but cause long-term maintenance debt. This skill implements a rigorous multi-layer audit that greps for literal values and ensures the dependency graph remains intact across the full stack.

Use cases

• Adding new subscription statuses that need to be reflected in Stripe webhooks, the DB, and the UI.
• Refactoring a legacy database column name without leaving behind stale type definitions.
• Standardizing user roles across an admin dashboard and backend authorization middleware.
• Ensuring seed scripts and test suites use authoritative constants instead of hardcoded strings.
• Known limitations: Assumes an existing constants file and migration-based schema workflow. Greenfield projects need these established first.