Skills

This skill is the agent's guarantee of production-grade, human-crafted backend engineering. It exists so that any backend the agent produces looks and behaves like a senior backend engineer built it on purpose for that specific product — properly validated, typed, secured, observable, and maintainable — never a throwaway tutorial scaffold. The agent must treat this skill as mandatory whenever it is triggered and complete the intent-first analysis below BEFORE writing any code. Skipping the analysis phase is forbidden.

When to use

Activate this skill when the user's prompt contains ANY of the following terms or signals:

• backend, API, server, server-side, service, endpoint, REST API, RESTful, CRUD, webhook, route.
• Express, Fastify, NestJS, Hono, Next.js API, Next.js route handler, Node.js, TypeScript server, tRPC.
• authentication, auth, JWT, OAuth, login, RBAC, authorization, password hashing, sessions.
• database, Prisma, Drizzle, ORM, PostgreSQL, MySQL, MongoDB, SQLite, schema, migration, Redis.
• production backend, production-grade, industry-level, secure API, best practice backend, scalable backend.
• build me a backend, build an API for my app, backend for my [web/mobile] app, add an API, make the server.
• validation, rate limiting, logging, observability, error handling, health check, file upload, queue, background job.
• User asks to implement an API contract, turn a requirement/PRD into a working backend, or generate framework-specific backend code.
• ANY task where the user expects working, production-worthy server code rather than an architecture discussion.

When in doubt, activate. This skill composes with backend-core (architecture design) and api-design (contracts): those decide WHAT and WHY; backend-craft decides HOW to implement it cleanly. requirements-analysis/PRD feeds the intent here.