google-labs-code/design.md

tdd-red-green-refactor

Enforces a disciplined Red-Green-Refactor (TDD) workflow in TypeScript/Node.js.

Quelltext ansehen

Verwendung

  1. Installationsbefehl kopieren und das Quellpaket vor der Installation in der eigenen Agentenumgebung prüfen.
  2. Benötigte Werkzeuge, Zugangsdaten und Hosts im Original prüfen.
  3. Eingabe und gewünschtes Ergebnis beschreiben und die Ausgabe mit den Quellen abgleichen.
npx skills add https://github.com/google-labs-code/design.md --skill tdd-red-green-refactor

Vor dem Start

  • Die Installation umfasst weder Modellabo noch API-Kontingent oder Dateizugriffsrechte.
  • Sterne und Installationen zeigen Aktivität, keine unabhängige Qualitäts- oder Sicherheitsprüfung.
Originales Skill-Dokument

Aus dem Quell-Repository gerendert; Überschriften, Beispiele, Code, Tabellen, Links und Bilder bleiben erhalten.

Red-Green-Refactor (TDD) Skill: TypeScript Edition

This skill implements a structural framework for AI-assisted programming to ensure every line of code is verifiable, typed, and purposeful.

The Three-Phase Cycle

Phase 1: Red (Establish Failure)

You must prove the feature does not exist and that your test is valid.

  1. Write One Test: Create a single test case (e.g., in Vitest or Jest) for the next small piece of behavior.
  2. Execute & Fail: Run the test. It must fail.
  3. Verify: Ensure the failure is related to the missing logic (e.g., ReferenceError: add is not defined) and not a configuration error.

Phase 2: Green (Minimal Pass)

Make the test pass as quickly and simply as possible.

  1. Minimal Implementation: Write the simplest code that satisfies the test. Do not build for the future; focus strictly on the current "Red" test.
  2. Run Tests: Execute the suite. All tests must be Green.
  3. Evidence: The transition from Red to Green is the "Proof of Work" for the developer.

Phase 3: Refactor (Clean Up)

Improve the code structure while maintaining the "Green" state.

  1. Clean Up: Improve naming, remove duplication, and optimize the code written in Phase 2.
  2. Safety Net: Rerun the tests after every change. If they turn Red, revert the change immediately.

Core Operational Rules

1. No "Horizontal Splurging"

You are strictly forbidden from writing a large "splurge" of multiple tests at once. You must follow a strictly incremental loop:

  • Write 1 Test -> See it Fail -> Write 1 Fix -> See it Pass.
  • Repeat this loop for every sub-feature.

2. Impose Backpressure

Use automated assertions and strong typing (TypeScript) as backpressure to prevent the AI from "guessing" the solution or "playing in the mud" with low-quality code.

3. Verification of Integrity

Never modify an existing test to make a failing implementation pass. If a test must change, it must be because the requirement changed, not because the code is difficult to write.


Example Workflow (TypeScript + Vitest)

Step 1: Red

typescript
// math.test.ts
import { describe, it, expect } from 'vitest';
import { add } from './math';

describe('add', () => {
  it('should sum two numbers', () => {
    expect(add(2, 2)).toBe(4); // Fails: ReferenceError: add is not defined
  });
});

Step 2: Green

typescript
// math.ts
export const add = (a: any, b: any) => {
  return 4; // Passes: Minimal code to satisfy the test
};

Step 3: Refactor

typescript
// math.ts
/**
 * Sums two numbers with explicit type safety.
 */
export const add = (a: number, b: number): number => {
  return a + b; // Passes: Proper implementation with safety net
};

Verwandte Aufgaben