api-graphql-designer_skill

Overview

This skill helps you design robust GraphQL schemas for standalone, federated, or stitched architectures. It generates SDL, resolver patterns, and production-ready guidance for pagination, DataLoader usage, and field-level security. Use it to move from domain models or REST endpoints to scalable GraphQL services with n+1 prevention and federation support.

How this skill works

I inspect provided entity definitions, schema_type, and optional federation_config to produce a complete SDL and resolver patterns. The skill applies tiered procedures: a fast T1 path for basic standalone schemas, T2 for production optimizations (DataLoader, cursor pagination, auth directives, federation v2), and T3 for advanced stitching, subscriptions, and monitoring. Outputs include schema_definition, resolver_patterns, optimization_config, and federation_config when required.

When to use it

• Designing a new GraphQL API from domain entities for a single service
• Converting REST endpoints into a unified GraphQL schema
• Implementing Apollo Federation v2 across microservices
• Preventing n+1 queries using DataLoader patterns
• Stitching multiple GraphQL services or integrating third-party schemas

Best practices

• Start with clear entity definitions and mark non-nullable fields based on business rules
• Use per-request DataLoader instances to batch and cache DB access for relationships
• Prefer cursor-based (Relay) pagination for large datasets and offset for simple admin lists
• Add field-level @auth directives and limit query complexity and depth (max 7–10)
• Define @key on federated entities and implement __resolveReference for each subgraph

Example use cases

• Generate a standalone SDL with Query/Mutation roots, input types, and offset pagination
• Produce a federated User subgraph with @key, reference resolvers, and DataLoader-backed relations
• Create a stitched gateway schema that delegates queries to legacy GraphQL services
• Add subscription definitions and PubSub patterns for real-time updates
• Prepare production config: query complexity limits, depth limiting, and caching strategy