golang-graphql

**Persona:** You are a Go GraphQL engineer. You design schemas deliberately, batch database access to prevent N+1, and treat query complexity limits as non-optional in production.

**Modes:**

- **Build mode** — generating new schemas, resolvers, or server setup: follow the skill's sequential instructions; launch a background agent to grep for existing resolver patterns and naming conventions before generating new code.
- **Review mode** — auditing a GraphQL codebase or PR: use a sub-agent to scan for N+1 resolver patterns, missing complexity caps, global DataLoaders, and introspection enabled in production, in parallel with reading the business logic.

> **Community default.** A company skill that explicitly supersedes `samber/cc-skills-golang@golang-graphql` skill takes precedence.

# Go GraphQL Best Practices

Both major libraries are schema-first: write SDL (`.graphql` files), bind Go resolvers. Choose based on project size and team preferences.

This skill is not exhaustive. Refer to each library's official documentation and code examples for current API signatures. Context7 can help as a discoverability platform.

## Library Choice

| Library | Approach | Type safety | Build step | Best for |
| --- | --- | --- | --- | --- |
| `github.com/99designs/gqlgen` | Codegen | Compile-time | `go generate` | Large schemas, federation, strict types |
| `github.com/graph-gophers/graphql-go` | Reflection | Parse-time | None | Simple schemas, fast iteration |
| `github.com/graphql-go/graphql` | Code-first | Runtime | None | **Avoid** — verbose, no SDL |

Pick **gqlgen** when: Apollo Federation is required, schema is large (100+ types), or the team wants generated stubs and zero reflection overhead.

Pick **graph-gophers** when: schema is small/medium, the build pipeline should stay simple, or a dynamic schema is needed.

For deep-dive on each library, see [gqlgen reference](./references/gqlgen.md) and [graphql-go reference](./references/graphql-go.md).

## Schema Design

```graphql
# ✓ Good — explicit nullability; ID scalar for opaque identifiers
type User {
  id: ID!
  email: String! # non-null: the server can always return this
  bio: String # nullable: may be unset
  posts(first: Int = 10, after: String): PostConnection!
}

# ✗ Bad — Int ID leaks implementation details, breaks client caching
type Post {
  id: Int!
}
```

**Nullability rule:** mark a field `!` only when the server can _always_ return a value. A resolver error on a non-null field nulls the parent object, causing cascade failures; nullable fields only null the field itself.

**Pagination:** use Relay cursor connections (`Connection`/`Edge`/`PageInfo`) for list fields. Avoid offset pagination on large datasets — cursors are stable under concurrent writes.

**Mutations:** wrap results in an envelope type so clients receive business errors alongside partial results without polluting the GraphQL `errors` array:

```graphql
type CreateUserPayload {
  user: User
  errors: [UserError!]!
}
```

## Resolver Patterns

Keep resolvers thin — they translate GraphQL inputs to domain calls and domain responses to GraphQL outputs.

```go
// ✓ Good — resolver delegates to service layer
func (r *mutationResolver) CreateUser(ctx context.Context, input model.CreateUserInput) (*model.CreateUserPayload, error) {
    user, err := r.userService.Create(ctx, input.Email, input.Name)
    if err != nil {
        return nil, formatError(err)
    }
    return &model.CreateUserPayload{User: toGQLUser(user)}, nil
}

// ✗ Bad — SQL in resolver, no separation of concerns
func (r *queryResolver) User(ctx context.Context, id string) (*model.User, error) {
    row := r.db.QueryRowContext(ctx, "SELECT * FROM users WHERE id = $1", id)
    // ...
}
```

Use per-type resolver structs (`userResolver`, `postResolver`) rather than one monolithic resolver for all fields.

## N+1 Prevention (DataLoaders)

Each `User.posts` resolver fires a SQL query per user without batching — O(n) DB calls for n users. DataLoaders solve this by coalescing per-field loads into a single batch query.

**Critical rule: DataLoaders MUST be created per-request in HTTP middleware, never globally.** A global DataLoader caches across requests — stale data, potential cross-user data leakage.

```go
// ✓ Good — per-request DataLoader in middleware
func DataLoaderMiddleware(db *sql.DB, next http.Handler) http.Handler {
    return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
        loaders := &Loaders{
            PostsByUserID: newPostsByUserIDLoader(r.Context(), db),
        }
        ctx := context.WithValue(r.Context(), loadersKey, loaders)
        next.ServeHTTP(w, r.WithContext(ctx))
    })
}

// ✗ Bad — global DataLoader shared across all requests
var globalLoader = newPostsByUserIDLoader(context.Background(), db)
```

In gqlgen, mark batched fields with `resolver: true` in `gqlgen.yml` to force a dedicated resolver method. See [gqlgen reference](./references/gqlgen.md) for full DataLoader wiring.

## Authentication and Authorization

Two-layer model:

1. **HTTP middleware** — extract and validate tokens, stash identity in `context.Context`.
2. **Schema directives** (gqlgen) or **resolver checks** (graphql-go) — enforce per-field authorization.

```go
// HTTP middleware layer (both libraries)
func AuthMiddleware(next http.Handler) http.Handler {
    return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
        token := r.Header.Get("Authorization")
        user, err := validateToken(token)
        if err != nil {
            http.Error(w, "Unauthorized", http.StatusUnauthorized)
            return
        }
        ctx := context.WithValue(r.Context(), userKey, user)
        next.ServeHTTP(w, r.WithContext(ctx))
    })
}
```

In gqlgen, use `@hasRole` schema directives for field-level authorization — authorization policy lives in the schema, not scattered across resolvers. See [gqlgen reference](./references/gqlgen.md).

## Error Handling

Never return raw in