safemocker solves the critical problem of testing next-safe-action v8 in Jest environments where ESM modules (.mjs files) cannot be directly imported. It provides a comprehensive mocking solution that's type-safe, robust, and extensive, allowing easy testing for all next-safe-action usage throughout your project.

📑 Table of Contents

• ✨ Features
• 🚀 Quick Start
• 📦 Installation
• 📚 Quick Start Guide
  • Jest Integration
  • Vitest Integration
• 📖 API Reference
  • Factory Functions
  • Configuration Options
• 💡 Usage Examples
  • Basic Action Testing
  • Validation Error Testing
  • Error Handling Testing
  • Custom Middleware Testing
  • Complex Integration Testing
  • Discriminated Unions & Complex Validation
  • Partial Updates & Batch Operations
• 🚀 Advanced Features
  • Nested Validation Errors
  • Discriminated Unions
  • Array Validation
  • Rate Limited Actions
• ⚙️ How It Works
  • Method Chaining
  • Middleware Chain Execution
  • SafeActionResult Structure
• 📁 Example Files
• ⚠️ Caveats & Considerations
• 🔧 Troubleshooting
• 🔄 Migration Guide
• 🤝 Contributing
• 🔗 Related Projects

✨ Features

• ✅ Works with Jest - Solves ESM compatibility issues (primary use case)
• ✅ Works with Vitest - Even with ESM support, mocking provides faster tests, easier control, consistent patterns, and better error scenario testing
• ✅ Replicates real middleware behavior - Auth, validation, error handling work exactly like the real library
• ✅ Returns proper SafeActionResult structure - Type-safe, matches real API exactly
• ✅ Type-safe API - Full TypeScript integration with proper inference
• ✅ Easy to use - Similar to Prismocker pattern, minimal setup required
• ✅ Standalone package - Can be extracted to separate repo for OSS distribution

🚀 Quick Start

import { createAuthedActionClient } from '@jsonbored/safemocker/jest';
import { z } from 'zod';

// Create authenticated action client
const authedAction = createAuthedActionClient();

// Define your action
const createUser = authedAction
  .inputSchema(z.object({ name: z.string().min(1), email: z.string().email() }))
  .metadata({ actionName: 'createUser', category: 'user' })
  .action(async ({ parsedInput, ctx }) => {
    return { id: 'new-id', ...parsedInput, createdBy: ctx.userId };
  });

// Test it!
const result = await createUser({ name: 'John', email: '[email protected]' });
expect(result.data).toEqual({ id: 'new-id', name: 'John', email: '[email protected]', createdBy: 'test-user-id' });

📦 Installation

npm install --save-dev @jsonbored/safemocker
# or
pnpm add -D @jsonbored/safemocker
# or
yarn add -D @jsonbored/safemocker

📚 Quick Start Guide

Step 1: Create Mock File

Create __mocks__/next-safe-action.ts in your project root:

import { createMockSafeActionClient } from '@jsonbored/safemocker/jest';

export const createSafeActionClient = createMockSafeActionClient({
  defaultServerError: 'Something went wrong',
  isProduction: false,
  auth: {
    enabled: true,
    testUserId: 'test-user-id',
    testUserEmail: '[email protected]',
    testAuthToken: 'test-token',
  },
});

export const DEFAULT_SERVER_ERROR_MESSAGE = 'Something went wrong';

Step 2: Use in Tests

// Your test file
import { authedAction } from './safe-action'; // Your real safe-action.ts file
import { z } from 'zod';

// Create action using REAL safe-action.ts (which uses mocked next-safe-action)
const testAction = authedAction
  .inputSchema(z.object({ id: z.string() }))
  .metadata({ actionName: 'testAction', category: 'user' })
  .action(async ({ parsedInput, ctx }) => {
    return { id: parsedInput.id, userId: ctx.userId };
  });

// Test SafeActionResult structure
const result = await testAction({ id: '123' });
expect(result.data).toEqual({ id: '123', userId: 'test-user-id' });
expect(result.serverError).toBeUndefined();
expect(result.fieldErrors).toBeUndefined();

Step 3: Verify Jest Auto-Mock

Jest will automatically use __mocks__/next-safe-action.ts when you import next-safe-action in your code. No additional configuration needed!

Step 1: Create Mock Setup

Create vitest.setup.ts or add to your test file:

import { vi } from 'vitest';
import { createMockSafeActionClient } from '@jsonbored/safemocker/vitest';

vi.mock('next-safe-action', () => {
  return {
    createSafeActionClient: createMockSafeActionClient({
      defaultServerError: 'Something went wrong',
      isProduction: false,
      auth: {
        enabled: true,
        testUserId: 'test-user-id',
        testUserEmail: '[email protected]',
        testAuthToken: 'test-token',
      },
    }),
    DEFAULT_SERVER_ERROR_MESSAGE: 'Something went wrong',
  };
});

Step 2: Use in Tests

// Your test file
import { authedAction } from './safe-action'; // Your real safe-action.ts file
import { z } from 'zod';

// Create action using REAL safe-action.ts (which uses mocked next-safe-action)
const testAction = authedAction
  .inputSchema(z.object({ id: z.string() }))
  .metadata({ actionName: 'testAction', category: 'user' })
  .action(async ({ parsedInput, ctx }) => {
    return { id: parsedInput.id, userId: ctx.userId };
  });

// Test SafeActionResult structure
const result = await testAction({ id: '123' });
expect(result.data).toEqual({ id: '123', userId: 'test-user-id' });
expect(result.serverError).toBeUndefined();
expect(result.fieldErrors).toBeUndefined();

Step 3: Configure Vitest

If using vitest.setup.ts, add it to your vitest.config.ts:

import { defineConfig } from 'vitest/config';

export default defineConfig({
  test: {
    setupFiles: ['./vitest.setup.ts'],
  },
});

📖 API Reference

Factory Functions

Creates a basic mock safe action client.

import { createMockSafeActionClient } from '@jsonbored/safemocker/jest'; // or 'safemocker/vitest'

const client = createMockSafeActionClient({
  defaultServerError: 'Something went wrong',
  isProduction: false,
  auth: {
    enabled: true,
    testUserId: 'test-user-id',
    testUserEmail: '[email protected]',
    testAuthToken: 'test-token',
  },
});

Creates a mock client with authentication middleware pre-configured.

import { createAuthedActionClient } from '@jsonbored/safemocker/jest';

const authedAction = createAuthedActionClient({
  auth: {
    testUserId: 'custom-user-id',
  },
});

const action = authedAction
  .inputSchema(z.object({ id: z.string() }))
  .action(async ({ parsedInput, ctx }) => {
    // ctx.userId, ctx.userEmail, ctx.authToken are available
    return { id: parsedInput.id, userId: ctx.userId };
  });

Creates a mock client with optional authentication middleware.

import { createOptionalAuthActionClient } from '@jsonbored/safemocker/jest';

const optionalAuthAction = createOptionalAuthActionClient();

const action = optionalAuthAction
  .inputSchema(z.object({ query: z.string() }))
  .action(async ({ parsedInput, ctx }) => {
    // ctx.user may be null, ctx.userId may be undefined
    return { query: parsedInput.query, userId: ctx.userId };
  });

Creates a mock client with rate limiting middleware.

import { createRateLimitedActionClient } from '@jsonbored/safemocker/jest';
import { z } from 'zod';

const metadataSchema = z.object({
  actionName: z.string().min(1),
  category: z.enum(['user', 'admin']).optional(),
});

const rateLimitedAction = createRateLimitedActionClient(metadataSchema);

const action = rateLimitedAction
  .inputSchema(z.object({ query: z.string() }))
  .metadata({ actionName: 'search', category: 'content' })
  .action(async ({ parsedInput }) => {
    return { results: [] };
  });

Creates all action client variants matching your real safe-action.ts pattern.

import { createCompleteActionClient } from '@jsonbored/safemocker/jest';
import { z } from 'zod';

const metadataSchema = z.object({
  actionName: z.string().min(1),
  category: z.enum(['user', 'admin', 'content']).optional(),
});

const {
  actionClient,
  loggedAction,
  rateLimitedAction,
  authedAction,
  optionalAuthAction,
} = createCompleteActionClient(metadataSchema, {
  auth: {
    testUserId: 'test-user-id',
  },
});

// Use exactly like your real safe-action.ts
const action = authedAction
  .inputSchema(z.object({ id: z.string() }))
  .metadata({ actionName: 'test' })
  .action(async ({ parsedInput, ctx }) => {
    return { id: parsedInput.id, userId: ctx.userId };
  });

Configuration Options

interface MockSafeActionClientConfig {
  defaultServerError?: string;        // Default: 'Something went wrong'
  isProduction?: boolean;              // Default: false
  auth?: {
    enabled?: boolean;                 // Default: true
    testUserId?: string;               // Default: 'test-user-id'
    testUserEmail?: string;            // Default: '[email protected]'
    testAuthToken?: string;            // Default: 'test-token'
  };
}

💡 Usage Examples

import { createAuthedActionClient } from '@jsonbored/safemocker/jest';
import { z } from 'zod';

const authedAction = createAuthedActionClient();

const createUser = authedAction
  .inputSchema(
    z.object({
      name: z.string().min(1),
      email: z.string().email(),
    })
  )
  .metadata({ actionName: 'createUser', category: 'user' })
  .action(async ({ parsedInput, ctx }) => {
    return {
      id: 'new-user-id',
      name: parsedInput.name,
      email: parsedInput.email,
      createdBy: ctx.userId,
    };
  });

// Test
const result = await createUser({
  name: 'John Doe',
  email: '[email protected]',
});

expect(result.data).toEqual({
  id: 'new-user-id',
  name: 'John Doe',
  email: '[email protected]',
  createdBy: 'test-user-id',
});

import { createAuthedActionClient } from '@jsonbored/safemocker/jest';
import { z } from 'zod';

const authedAction = createAuthedActionClient();

const updateProfile = authedAction
  .inputSchema(
    z.object({
      name: z.string().min(1),
      email: z.string().email(),
    })
  )
  .action(async ({ parsedInput }) => {
    return { success: true };
  });

// Test validation errors
const result = await updateProfile({
  name: '', // Invalid: min length
  email: 'invalid-email', // Invalid: not an email
});

expect(result.fieldErrors).toBeDefined();
expect(result.fieldErrors?.name).toBeDefined();
expect(result.fieldErrors?.email).toBeDefined();
expect(result.data).toBeUndefined();

import { createAuthedActionClient } from '@jsonbored/safemocker/jest';
import { z } from 'zod';

const authedAction = createAuthedActionClient({
  defaultServerError: 'Something went wrong',
  isProduction: false, // Use error message in development
});

const deleteItem = authedAction
  .inputSchema(z.object({ id: z.string() }))
  .action(async () => {
    throw new Error('Item not found');
  });

// Test error handling
const result = await deleteItem({ id: 'test-id' });

expect(result.serverError).toBe('Item not found');
expect(result.data).toBeUndefined();

// Test production mode (hides error details)
const prodAction = createAuthedActionClient({
  defaultServerError: 'Something went wrong',
  isProduction: true,
});

const prodResult = await prodAction
  .inputSchema(z.object({ id: z.string() }))
  .action(async () => {
    throw new Error('Sensitive error details');
  })({ id: 'test' });

expect(prodResult.serverError).toBe('Something went wrong');
expect(prodResult.serverError).not.toBe('Sensitive error details');

import { createMockSafeActionClient } from '@jsonbored/safemocker/jest';
import { z } from 'zod';

const client = createMockSafeActionClient();

// Add custom middleware
client.use(async ({ next, ctx = {} }) => {
  // Add custom context (next-safe-action format: { ctx: newContext })
  return next({ ctx: { ...ctx, customValue: 'test' } });
});

const action = client
  .inputSchema(z.object({ id: z.string() }))
  .action(async ({ parsedInput, ctx }) => {
    return {
      id: parsedInput.id,
      customValue: ctx.customValue,
    };
  });

const result = await action({ id: 'test-id' });

expect(result.data).toEqual({
  id: 'test-id',
  customValue: 'test',
});

import { createCompleteActionClient } from '@jsonbored/safemocker/jest';
import { z } from 'zod';

const metadataSchema = z.object({
  actionName: z.string().min(1),
  category: z.enum(['user', 'admin']).optional(),
});

const { authedAction } = createCompleteActionClient(metadataSchema, {
  auth: {
    testUserId: 'user-123',
    testUserEmail: '[email protected]',
  },
});

// Replicate your real safe-action.ts pattern
const updateJob = authedAction
  .inputSchema(
    z.object({
      jobId: z.string().uuid(),
      title: z.string().min(1),
      status: z.enum(['draft', 'published', 'archived']),
    })
  )
  .metadata({ actionName: 'updateJob', category: 'user' })
  .action(async ({ parsedInput, ctx }) => {
    // Verify context is injected
    expect(ctx.userId).toBe('user-123');
    expect(ctx.userEmail).toBe('[email protected]');

    return {
      jobId: parsedInput.jobId,
      title: parsedInput.title,
      status: parsedInput.status,
      updatedBy: ctx.userId,
    };
  });

// Test success case
const successResult = await updateJob({
  jobId: '123e4567-e89b-12d3-a456-426614174000',
  title: 'Software Engineer',
  status: 'published',
});

expect(successResult.data).toEqual({
  jobId: '123e4567-e89b-12d3-a456-426614174000',
  title: 'Software Engineer',
  status: 'published',
  updatedBy: 'user-123',
});

// Test validation errors
const validationResult = await updateJob({
  jobId: 'invalid-uuid',
  title: '',
  status: 'invalid-status',
});

expect(validationResult.fieldErrors).toBeDefined();
expect(validationResult.fieldErrors?.jobId).toBeDefined();
expect(validationResult.fieldErrors?.title).toBeDefined();
expect(validationResult.fieldErrors?.status).toBeDefined();

import { createAuthedActionClient } from '@jsonbored/safemocker/jest';
import { z } from 'zod';

const authedAction = createAuthedActionClient();

// Discriminated union for content types
const articleSchema = z.object({
  type: z.literal('article'),
  title: z.string().min(1),
  content: z.string().min(1),
  author: z.string().min(1),
});

const videoSchema = z.object({
  type: z.literal('video'),
  title: z.string().min(1),
  videoUrl: z.string().url(),
  duration: z.number().int().positive(),
});

const contentSchema = z.discriminatedUnion('type', [articleSchema, videoSchema]);

const createContent = authedAction
  .inputSchema(
    z.object({
      content: contentSchema,
      category: z.enum(['tech', 'business']),
    })
  )
  .action(async ({ parsedInput, ctx }) => {
    return {
      id: 'content-1',
      ...parsedInput.content,
      category: parsedInput.category,
      createdBy: ctx.userId,
    };
  });

// Test article content
const articleResult = await createContent({
  content: {
    type: 'article',
    title: 'Test Article',
    content: 'Article content...',
    author: 'John Doe',
  },
  category: 'tech',
});

expect(articleResult.data?.type).toBe('article');

// Test video content
const videoResult = await createContent({
  content: {
    type: 'video',
    title: 'Test Video',
    videoUrl: 'https://example.com/video.mp4',
    duration: 300,
  },
  category: 'tech',
});

expect(videoResult.data?.type).toBe('video');

// Test validation errors (nested fields use dot notation)
const invalidResult = await createContent({
  content: {
    type: 'article',
    title: '', // Invalid
    content: '', // Invalid
    author: '', // Invalid
  } as any,
  category: 'tech',
});

expect(invalidResult.fieldErrors?.['content.title']).toBeDefined();
expect(invalidResult.fieldErrors?.['content.content']).toBeDefined();
expect(invalidResult.fieldErrors?.['content.author']).toBeDefined();

import { createAuthedActionClient } from '@jsonbored/safemocker/jest';
import { z } from 'zod';

const authedAction = createAuthedActionClient();

// Partial update action
const updateContent = authedAction
  .inputSchema(
    z.object({
      contentId: z.string().uuid(),
      updates: z.object({
        title: z.string().min(1).optional(),
        published: z.boolean().optional(),
        tags: z.array(z.string()).max(10).optional(),
      }),
    })
  )
  .action(async ({ parsedInput, ctx }) => {
    return {
      id: parsedInput.contentId,
      updatedFields: Object.keys(parsedInput.updates),
      updatedBy: ctx.userId,
    };
  });

// Test partial update
const result = await updateContent({
  contentId: '123e4567-e89b-12d3-a456-426614174000',
  updates: {
    title: 'Updated Title',
    published: true,
  },
});

expect(result.data?.updatedFields).toContain('title');
expect(result.data?.updatedFields).toContain('published');

// Batch update action
const batchUpdate = authedAction
  .inputSchema(
    z.object({
      updates: z.array(
        z.object({
          contentId: z.string().uuid(),
          updates: z.object({
            title: z.string().min(1).optional(),
          }),
        })
      ).min(1).max(50),
    })
  )
  .action(async ({ parsedInput }) => {
    return {
      totalUpdated: parsedInput.updates.length,
      updated: parsedInput.updates.map((u) => u.contentId),
    };
  });

// Test batch update
const batchResult = await batchUpdate({
  updates: [
    { contentId: 'id-1', updates: { title: 'Title 1' } },
    { contentId: 'id-2', updates: { title: 'Title 2' } },
  ],
});

expect(batchResult.data?.totalUpdated).toBe(2);

🚀 Advanced Features

When using nested objects in your schemas, validation errors use dot notation for field paths:

const schema = z.object({
  content: z.object({
    title: z.string().min(1),
    author: z.string().min(1),
  }),
});

// Invalid input
const result = await action({
  content: {
    title: '', // Invalid
    author: '', // Invalid
  },
});

// Field errors use dot notation
expect(result.fieldErrors?.['content.title']).toBeDefined();
expect(result.fieldErrors?.['content.author']).toBeDefined();

safemocker fully supports Zod discriminated unions:

const contentSchema = z.discriminatedUnion('type', [
  z.object({ type: z.literal('article'), content: z.string() }),
  z.object({ type: z.literal('video'), videoUrl: z.string().url() }),
]);

const action = client
  .inputSchema(z.object({ content: contentSchema }))
  .action(async ({ parsedInput }) => {
    // TypeScript knows the discriminated union type
    if (parsedInput.content.type === 'article') {
      // parsedInput.content.content is available
    } else {
      // parsedInput.content.videoUrl is available
    }
  });

Complex array validation with nested items:

const schema = z.object({
  items: z.array(
    z.object({
      id: z.string().uuid(),
      name: z.string().min(1),
    })
  ).min(1).max(50),
});

// Validation errors for arrays
const result = await action({ items: [] }); // Invalid: min 1
expect(result.fieldErrors?.items).toBeDefined();

Rate limiting middleware is included in rateLimitedAction:

import { createRateLimitedActionClient } from '@jsonbored/safemocker/jest';
import { z } from 'zod';

const metadataSchema = z.object({
  actionName: z.string().min(1),
  category: z.enum(['content', 'user']).optional(),
});

const rateLimitedAction = createRateLimitedActionClient(metadataSchema);

const searchAction = rateLimitedAction
  .inputSchema(z.object({ query: z.string() }))
  .metadata({ actionName: 'search', category: 'content' })
  .action(async ({ parsedInput }) => {
    return { results: [] };
  });

⚙️ How It Works

safemocker replicates the exact method chaining pattern of next-safe-action:

client
  .inputSchema(zodSchema)      // Step 1: Define input validation
  .metadata(metadata)          // Step 2: Add metadata (optional)
  .action(handler)              // Step 3: Define action handler

1. Input Validation - Zod schema validation happens first
2. Middleware Execution - Middleware runs in order, each can modify context
3. Handler Execution - Action handler runs with validated input and context
4. Result Wrapping - Handler result is wrapped in SafeActionResult structure
5. Error Handling - Any errors are caught and converted to serverError

All actions return a SafeActionResult<TData>:

interface SafeActionResult<TData> {
  data?: TData;                              // Success data
  serverError?: string;                       // Server error message
  fieldErrors?: Record<string, string[]>;     // Validation errors by field
  validationErrors?: Record<string, string[]>; // General validation errors
}

📁 Example Files

The safemocker package includes comprehensive examples with full test coverage:

Real-world safe-action.ts pattern using mocked next-safe-action. Demonstrates:

• Base action client creation
• Metadata schema definition
• Error handling configuration
• Middleware chaining
• Complete action client factory pattern

Test Coverage: __tests__/real-integration.test.ts (comprehensive integration tests)

User management actions demonstrating common patterns:

• createUser - Authentication required, input validation, context injection
• getUserProfile - Optional authentication, UUID validation
• updateUserSettings - Partial updates, enum validation
• deleteUser - Admin-only pattern, UUID validation

Test Coverage: __tests__/real-integration.test.ts

Complex content management actions demonstrating advanced v8 features:

• createContent - Discriminated unions (article, video, podcast), nested validation, array constraints
• updateContent - Partial updates with optional fields, UUID validation
• batchUpdateContent - Array validation with complex items, batch operations
• searchContent - Rate limiting, complex query validation, pagination, filtering
• getContentWithRelations - Optional authentication, nested relations, conditional data

Features Demonstrated:

• ✅ Discriminated unions with type-specific validation
• ✅ Nested object validation (dot notation for errors)
• ✅ Array validation with min/max constraints
• ✅ Partial updates with optional fields
• ✅ Batch operations with array validation
• ✅ Complex query parameters with defaults
• ✅ Rate limiting middleware
• ✅ Optional authentication with conditional logic
• ✅ Nested relations and conditional data inclusion

Test Coverage: __tests__/content-actions.test.ts (30 comprehensive tests, all passing)

⚠️ Caveats & Considerations

Problem: Jest cannot directly import ESM modules (.mjs files) without experimental configuration. next-safe-action is ESM-only.

Solution: safemocker provides a CommonJS-compatible mock that Jest can import directly. Your real safe-action.ts file uses the mocked next-safe-action, so you test the real middleware logic with mocked dependencies.

Important: Always use your real safe-action.ts file in tests. Don't mock it - mock next-safe-action instead.

Real next-safe-action middleware:

• Executes in actual Next.js server environment
• Has access to headers(), cookies(), etc.
• Performs real authentication checks
• Makes real database calls

safemocker middleware:

• Executes in test environment
• Uses test configuration (test user IDs, etc.)
• Skips real authentication (injects test context)
• No real database calls

Key Point: The middleware logic is replicated, but the implementation uses test-friendly mocks. This allows you to test your action handlers with realistic middleware behavior without needing a full Next.js server environment.

safemocker maintains full type safety:

• ✅ Input schemas are type-checked
• ✅ Handler parameters are typed (parsedInput, ctx)
• ✅ Return types are inferred
• ✅ SafeActionResult is properly typed

Note: TypeScript may show errors in your IDE if next-safe-action types aren't available. This is expected - the runtime behavior is correct, and types are provided by safemocker.

Development Mode (isProduction: false):

• Error messages include full details
• Useful for debugging during development

Production Mode (isProduction: true):

• Error messages use defaultServerError
• Hides sensitive error details
• Matches real next-safe-action behavior

Recommendation: Use isProduction: false in tests to see actual error messages, but test both modes to ensure your error handling works correctly.

Default Behavior:

• Authentication is always successful in tests
• Test user context is always injected
• No real authentication checks are performed

Why: In tests, you want to focus on testing your action logic, not authentication infrastructure. Real authentication should be tested separately with integration tests.

Customization:

• Set auth.enabled: false to disable auth middleware
• Set custom testUserId, testUserEmail, testAuthToken for different test scenarios
• Use different auth configs for different test suites

Real next-safe-action:

• Metadata validation happens in middleware
• Invalid metadata throws errors

safemocker:

• Metadata validation is replicated
• Use createMetadataValidatedActionClient() or createCompleteActionClient() with metadata schema
• Invalid metadata throws 'Invalid action metadata' error

Recommendation: Always provide metadata in tests to match real usage patterns.

🔧 Troubleshooting

Problem: Jest cannot find the next-safe-action module.

Solution: Ensure your __mocks__/next-safe-action.ts file is in the correct location (project root or __mocks__ directory at package level).

Verify:

# Check mock file exists
ls __mocks__/next-safe-action.ts

# Check Jest is using the mock
# Add console.log in your mock file to verify it's being loaded

Problem: Vitest isn't using the mock.

Solution: Ensure vi.mock('next-safe-action', ...) is called before any imports that use next-safe-action.

Best Practice: Put mock setup in vitest.setup.ts or at the top of your test file before any imports.

Problem: TypeScript shows errors about missing exports from next-safe-action.

Solution: This is expected - safemocker provides runtime mocks, but TypeScript may not recognize them. The code will work correctly at runtime.

Workaround: Add type assertions if needed, but the runtime behavior is correct.

Problem: ctx.userId or other context values are undefined.

Solution: Ensure you're using authedAction or optionalAuthAction (not base actionClient), and that auth.enabled is true in config.

Check:

const client = createAuthedActionClient({
  auth: {
    enabled: true, // Must be true
    testUserId: 'test-user-id',
  },
});

Problem: Invalid input doesn't return fieldErrors.

Solution: Ensure you're using .inputSchema() with a Zod schema. Validation happens automatically.

Verify:

const action = client
  .inputSchema(z.object({ email: z.string().email() })) // Schema required
  .action(async ({ parsedInput }) => { ... });

const result = await action({ email: 'invalid' });
expect(result.fieldErrors).toBeDefined(); // Should have email error

Problem: Testing nested validation errors but not finding them.

Solution: Nested fields use dot notation in fieldErrors.

Example:

const schema = z.object({
  user: z.object({
    email: z.string().email(),
  }),
});

const result = await action({ user: { email: 'invalid' } });
// Use dot notation:
expect(result.fieldErrors?.['user.email']).toBeDefined();
// NOT: result.fieldErrors?.user?.email

🔄 Migration Guide

Before (Manual Mock):

vi.mock('./safe-action.ts', async () => {
  const createActionHandler = (inputSchema: any) => {
    return vi.fn((handler: any) => {
      return async (input: unknown) => {
        try {
          const parsed = inputSchema ? inputSchema.parse(input) : input;
          const result = await handler({
            parsedInput: parsed,
            ctx: { userId: 'test-user-id' },
          });
          return result;
        } catch (error) {
          throw error;
        }
      };
    });
  };
  // ... complex mock setup
});

After (safemocker):

// __mocks__/next-safe-action.ts
import { createMockSafeActionClient } from '@jsonbored/safemocker/jest';

export const createSafeActionClient = createMockSafeActionClient({
  auth: { testUserId: 'test-user-id' },
});

// Use REAL safe-action.ts in tests
import { authedAction } from './safe-action';

Benefits:

• ✅ Less boilerplate
• ✅ Consistent SafeActionResult structure
• ✅ Real middleware behavior replication
• ✅ Type-safe
• ✅ Easier to maintain

🤝 Contributing

This package is designed to be standalone and extractable. Contributions welcome!

License

MIT

Author

JSONbored

🔗 Related Projects

• next-safe-action - The real library being mocked
• Prismocker - Similar type-safe mocking tool for Prisma Client (inspiration for this package)
• Claude Pro Directory - The parent project where safemocker and prismocker were originally developed