Testing with Claude Code¶

Learn how to write, run, and debug tests with Claude Code.

What You'll Learn¶

Writing unit tests
Running test suites
Debugging failing tests
Test coverage strategies

Running Tests¶

Basic Test Run¶

> Run the tests

Claude identifies your test framework and runs tests: - npm test - pytest - go test - cargo test - etc.

Specific Tests¶

> Run only the auth tests

> Run the test for getUserById function

> Run tests matching "login"

With Options¶

> Run tests with verbose output

> Run tests and show coverage

> Run tests in watch mode

Understanding Test Output¶

When Tests Pass¶

> Run the tests

✓ All 42 tests passed in 3.2s

When Tests Fail¶

> Run the tests

  FAIL src/auth.test.js
    ✕ should validate email format (12ms)

    Expected: true
    Received: false

    at validateEmail (src/utils.js:15:10)

Claude can then help:

> Why is the email validation test failing?

Writing Tests¶

For Existing Code¶

> Write tests for the UserService class

Claude: 1. Reads the class 2. Identifies testable functions 3. Generates appropriate test cases

With Specifications¶

> Write tests for validatePassword that check:
  - Minimum 8 characters
  - At least one uppercase
  - At least one number
  - Common passwords rejected

Following Existing Patterns¶

> Write tests for ProductService following the same
  pattern as UserService.test.js

Test Types¶

Unit Tests¶

> Write unit tests for the calculateDiscount function

Tests isolated logic with mocked dependencies.

Integration Tests¶

> Write an integration test for the checkout flow

Tests multiple components working together.

End-to-End Tests¶

> Write an E2E test for user registration

Tests the full user journey.

Test-Driven Development¶

Claude supports TDD workflows:

Step 1: Write the Test First¶

> Write a test for a function called `slugify` that:
  - Converts "Hello World" to "hello-world"
  - Removes special characters
  - Handles empty strings

Step 2: Implement the Function¶

> Now implement slugify to pass those tests

Step 3: Run and Verify¶

> Run the slugify tests

Debugging Failing Tests¶

Understand the Failure¶

> This test is failing:

  expect(result).toEqual({ name: 'John', age: 30 });

  Received: { name: 'John', age: '30' }

  What's wrong?

Claude identifies the type mismatch.

Trace the Issue¶

> The test passes locally but fails in CI. Help me understand why.

Fix the Test vs Fix the Code¶

> Is the test wrong or is the code wrong?

Claude analyzes the expected behavior.

Coverage¶

Check Coverage¶

> What's the test coverage for the auth module?

Improve Coverage¶

> Which functions in utils.js are not tested?

> Add tests to cover the edge cases in formatDate

Coverage Goals¶

> What would it take to get auth.js to 80% coverage?

Common Testing Scenarios¶

Async Functions¶

> Write tests for this async function that fetches user data

Claude handles: - async/await - Promises - Mocking fetch calls

Error Handling¶

> Add tests that verify the function throws correct errors

Edge Cases¶

> What edge cases should I test for parsePhoneNumber?

Claude suggests: - Empty input - Invalid formats - International numbers - Special characters

Mocking¶

> Write tests for sendEmail that mock the email service

Framework-Specific Help¶

Jest¶

> Set up Jest for this project

> How do I mock a module in Jest?

> Use Jest snapshots to test the Header component

Pytest¶

> Use pytest fixtures for database setup

> How do I parametrize this test?

React Testing Library¶

> Test that clicking the button triggers the callback

> Test that the loading state shows a spinner

Test Maintenance¶

Update Tests After Changes¶

> I changed the User model to include 'role'. Update the tests.

Remove Obsolete Tests¶

> Which tests are for deleted functionality?

Improve Test Quality¶

> Review the auth tests for flakiness and bad practices

Try It Yourself¶

Exercise: Test-Driven Development¶

Start Claude Code in a new directory

Write tests first:

> Write tests for a function called `isPrime` that:
  - Returns true for prime numbers (2, 3, 5, 7, 11)
  - Returns false for non-primes (1, 4, 6, 8, 9, 10)
  - Throws for negative numbers and non-integers

Run tests (they fail):
```
> Run the tests
```
Implement the function:
```
> Implement isPrime to pass the tests
```
Run again:
```
> Run the tests
```

Exercise: Coverage Challenge¶

Take an existing file with low coverage

Ask Claude what's not tested:

> What code paths in auth.js have no tests?

Write tests for uncovered paths
Verify coverage improved

Pro Tips¶

Test Names Matter¶

> Rename these tests to describe the behavior being tested,
  not just the function name

Avoid Testing Implementation¶

> This test is brittle because it tests implementation details.
  Rewrite it to test behavior.

One Assertion Per Test¶

> Split this test into multiple tests, one for each assertion

Use Before/After Hooks¶

> Set up common test fixtures using beforeEach

What's Next?¶

You've completed intermediate tutorials! Move on to Advanced: Custom Commands to extend Claude Code.

Summary: - Claude can run tests and interpret results - Generate tests based on code analysis or specifications - Support for TDD: write tests first, then implement - Debugging failing tests by analyzing expected vs actual - Coverage analysis and improvement suggestions

Learning Resources¶

Featured Video¶

GitHub: TDD with GitHub Copilot (Official GitHub channel)

Test-Driven Development with AI - write tests first, implement, verify with AI assistance.

Additional Resources¶

Type	Resource	Description
🎬 Video	Claude Code Testing Tips	Edmund Yong - TDD workflows
📚 Official Docs	GitHub TDD Guide	Official TDD documentation
📖 Tutorial	TDD with AI	Builder.io AI TDD patterns
🎓 Free Course	Microsoft TDD Course	Free paired programming course
💼 Commercial	TDD Complete Course	Comprehensive TDD training