Building GraphQL Server

building-graphql-server

Build production-ready GraphQL servers with schema design, resolvers,

jeremylongshore/claude-code-plugins-plus-skills·0installs2.8Kstars

SKILL.md

Full skill instructions

Building GraphQL Server

Overview

Build production-ready GraphQL servers with SDL-first or code-first schema design, efficient resolver implementations with DataLoader batching, real-time subscriptions via WebSocket, and field-level authorization. Support Apollo Server, Yoga, Mercurius, and Strawberry across Node.js and Python runtimes.

Prerequisites

Node.js 18+ with Apollo Server/Yoga/Mercurius, or Python 3.10+ with Strawberry/Ariadne
Database with ORM (Prisma, TypeORM, SQLAlchemy) for resolver data sources
Redis for subscription pub/sub and DataLoader caching (production deployments)
GraphQL client for testing: GraphiQL, Apollo Studio, or Insomnia
graphql-codegen for TypeScript type generation from schema (recommended)

Instructions

Examine existing data models, database schemas, and business requirements using Read and Glob to determine the entity graph and relationship structure.
Design the GraphQL schema with type definitions, including Query, Mutation, and Subscription root types, input types for mutations, and connection types for paginated lists.
Implement resolvers for each field, using DataLoader to batch and deduplicate database queries for nested relationships (N+1 query prevention).
Add input validation on mutation arguments using custom scalars (DateTime, Email, URL) and directive-based validation (@constraint(minLength: 1, maxLength: 255)).
Implement field-level authorization using schema directives (@auth(requires: ADMIN)) or resolver middleware that checks user roles from the GraphQL context.
Configure query complexity analysis and depth limiting to prevent abusive queries (maximum depth of 7, maximum complexity score of 1000).
Set up real-time subscriptions using graphql-ws protocol over WebSocket with Redis pub/sub for multi-instance message distribution.
Generate TypeScript types from the schema using graphql-codegen to ensure type safety between schema definitions and resolver implementations.
Write integration tests using executeOperation for query/mutation testing and WebSocket client tests for subscription verification.

See ${CLAUDE_SKILL_DIR}/references/implementation.md for the full implementation guide.

Output

${CLAUDE_SKILL_DIR}/src/schema/ - GraphQL SDL type definitions organized by domain
${CLAUDE_SKILL_DIR}/src/resolvers/ - Resolver implementations per type with DataLoader integration
${CLAUDE_SKILL_DIR}/src/dataloaders/ - DataLoader factories for batched database queries
${CLAUDE_SKILL_DIR}/src/directives/ - Custom schema directives (auth, validation, caching)
${CLAUDE_SKILL_DIR}/src/scalars/ - Custom scalar type definitions (DateTime, JSON, Email)
${CLAUDE_SKILL_DIR}/src/subscriptions/ - Subscription resolvers with pub/sub configuration
${CLAUDE_SKILL_DIR}/generated/types.ts - Auto-generated TypeScript types from schema

Error Handling

Error	Cause	Solution
N+1 query detected	Resolver fetches related records individually inside list resolver	Wrap data access in DataLoader; batch by parent ID array; cache within request scope
Query complexity exceeded	Client sends deeply nested query exceeding complexity budget	Return error with current complexity score and maximum allowed; suggest query simplification
Subscription connection dropped	WebSocket heartbeat timeout or network interruption	Implement automatic reconnection in client; use `graphql-ws` `connectionInitWaitTimeout`
Partial resolver failure	One field resolver throws while others succeed	Return partial data with `errors` array per GraphQL spec; log failed resolver with context
Schema stitching conflict	Duplicate type names when merging multiple schema modules	Use schema namespacing or federation with `@key` directives to resolve type ownership

Refer to ${CLAUDE_SKILL_DIR}/references/errors.md for comprehensive error patterns.

Examples

E-commerce product catalog: Schema with Product, Category, Review types, DataLoader-backed resolvers for nested queries like products { reviews { author } }, and a productUpdated subscription for inventory changes.

Multi-tenant SaaS dashboard: Code-first schema using TypeGraphQL decorators, tenant-scoped resolvers extracting tenantId from JWT context, and field-level visibility based on subscription plan tier.

Federated microservice graph: Apollo Federation with @key and @external directives across User, Order, and Product subgraphs, composed into a unified supergraph with a gateway router.

See ${CLAUDE_SKILL_DIR}/references/examples.md for additional examples.

Resources

GraphQL Specification: https://spec.graphql.org/
Apollo Server documentation: https://www.apollographql.com/docs/apollo-server/
DataLoader pattern: https://github.com/graphql/dataloader
graphql-ws protocol: https://github.com/enisdenjo/graphql-ws

Install skill

Copy and run in your terminal

npx -y skills add https://github.com/jeremylongshore/claude-code-plugins-plus-skills --skill building-graphql-server

View on GitHub

plugins/api-development/graphql-server-builder/skills/building-graphql-server/SKILL.md

planning-disaster-recovery
Execute use when you need to work with backup and recovery.
1 installs
exa-core-workflow-b
Execute Exa findSimilar, getContents, answer, and streaming answer workflows.
0 installs
scanning-api-security
Detect API security vulnerabilities including injection, broken auth,
0 installs
mistral-cost-tuning
Optimize Mistral AI costs through model selection, token management,
0 installs