Typescript Standard

TypeScript & Testing Standards

Language-specific rules for TypeScript projects covering clean code constraints, TDD workflow, and testing framework conventions.

When to Activate

• Writing or modifying TypeScript code
• Creating or updating tests in a TypeScript project
• Reviewing TypeScript code for standards compliance
• Setting up testing infrastructure (Jest, Playwright)

Core Concepts

• Strict typing everywhere — no any, interfaces for all data structures
• Small methods — 10-20 lines maximum, extraction is mandatory
• TDD flow — one test at a time, written before the implementation

Detailed Guidance

Clean Code Constraints

• Method Size: Methods must be between 10 and 20 lines max.
• Typing: Strict mode enabled. No any. Use interfaces for all data structures.
• Naming: Variable and function names must be descriptive and self-documenting.
• Simplification: Avoid unnecessary abstractions. Prefer the simplest working solution.

Testing Protocol (TDD)

• One Test at a Time: Create exactly one test before implementing the corresponding code.
• Given-When-Then: Use this format in comments to structure test logic.
• Data Builders: Use the entity.data-builder.ts pattern for test data.
• Faker: Use Faker for all default values to avoid hardcoded bias.
• Naming Convention: Use file-name.spec.ts and place tests near the production code.

Testing Frameworks

• Unit/Component: Jest + React Testing Library.
• E2E: Playwright
• Mocks: Mock database calls; use InMemoryRepositories for adapters.
• Coverage: Minimum 80% coverage required.

Examples

Test structure:

// file: create-user.spec.ts
describe('CreateUser', () => {
  it('should create a user with valid data', () => {
    // Given
    const userData = aUser().withEmail('test@example.com').build();
    // When
    const result = createUser(userData);
    // Then
    expect(result.email).toBe('test@example.com');
  });
});

Guidelines

1. Strict mode enabled — zero tolerance for any
2. Methods between 10-20 lines; extract if longer
3. One test before one implementation — no batch writing
4. Use Given-When-Then comments in every test
5. Data builders with Faker for test data — no hardcoded values
6. Tests live next to production code as *.spec.ts
7. Minimum 80% test coverage

Integration

• Related: recursive-exploration (reasoning model), refactor (cleanup workflow), code-review (review checklist)

Skill Metadata

• Created: 2025-06-01
• Last Updated: 2025-07-01
• Author: didacrios
• Version: 1.0.0