QE as AI Orchestrator
Lab Playbook 1

Your local folder and test_case.md are ready. The agent will be able to read this file directly from your filesystem in the next phases.

I'll set up your local training workspace in sequence.

**Step 1 — Initialize local repository**
```bash
$ git init
Initialized empty Git repository in /your-workspace/training-folder/.git/
✅ Git repository initialized.
```

**Step 2 — Create mock project structure**
I've created the following files to simulate our target application:
- `package.json`
- `src/pages/login.tsx`
- `src/pages/dashboard.tsx`
- `tests/` directory

**Step 3 — Initial commit**
```bash
$ git add .
$ git commit -m "chore: initial project setup"
[main (root-commit) 4a69361] chore: initial project setup
 4 files changed, 19 insertions(+)
 create mode 100644 package.json
 create mode 100644 src/pages/dashboard.tsx
 create mode 100644 src/pages/login.tsx
 create mode 100644 test_case.md
✅ Initial commit complete.
```

**Step 4 — Verify directory structure**
```
├── .git/
├── src/
│   ├── pages/
│   │   └── login.tsx
│   │   └── dashboard.tsx
├── package.json
├── test_case.md
└── tests/
```

✅ All steps complete. Local project is successfully initialized and tracked by Git. Ready for your next instruction.

I've read test_case.md. Here is my structured understanding:

**Core Requirements (Critical Steps)**
1. User must be able to navigate to /login
2. User must be able to submit valid credentials (email + password)
3. System must redirect to /dashboard on successful login

**Bolded Critical Steps (Non-Negotiable Test Gates)**
- ✅ Navigate to /login
- ✅ Enter valid credentials
- ✅ Redirect to /dashboard on success

**Acceptance Criteria as Testable Assertions**
| Input | Expected Output |
|-------|----------------|
| Valid email + correct password | Redirect to /dashboard |
| Valid email + wrong password | Error: "Invalid credentials" |
| Empty email field | Error: "Email is required" |
| Empty password field | Error: "Password is required" |
| Successful login | Session cookie persisted |

**Non-Functional Edge Cases**
- Performance: Login response must be < 2 seconds (requires timer assertion)
- Accessibility: ARIA labels needed on all input fields (requires a11y checks)

✅ Understanding confirmed. Awaiting your instruction to begin automation.

Creating ./tests/login.spec.ts now...

```typescript
import { test, expect } from '@playwright/test';

test.describe('Login Feature — QE Runbook Lab 1', () => {

  // ── Critical Step 1: Navigate to /login ──────────────────
  test('should load the login page', async ({ page }) => {
    const start = Date.now();
    await page.goto('/login');
    const duration = Date.now() - start;

    await expect(page).toHaveURL('/login');
    expect(duration).toBeLessThan(2000); // Non-functional: < 2s
  });

  // ── Critical Step 2 & 3: Valid login → /dashboard ────────
  test('valid credentials redirect to /dashboard', async ({ page, context }) => {
    await page.goto('/login');
    await page.fill('[aria-label="Email"]', 'user@example.com');
    await page.fill('[aria-label="Password"]', 'SecurePass123');
    await page.click('[type="submit"]');

    await expect(page).toHaveURL('/dashboard');

    // Session cookie assertion
    const cookies = await context.cookies();
    const session = cookies.find(c => c.name === 'session');
    expect(session).toBeDefined();
  });

  // ── Acceptance: wrong password → error message ────────────
  test('wrong password shows Invalid credentials error', async ({ page }) => {
    await page.goto('/login');
    await page.fill('[aria-label="Email"]', 'user@example.com');
    await page.fill('[aria-label="Password"]', 'WrongPass');
    await page.click('[type="submit"]');

    await expect(page.getByText('Invalid credentials')).toBeVisible();
  });

  // ── Acceptance: empty email → validation error ────────────
  test('empty email shows Email is required error', async ({ page }) => {
    await page.goto('/login');
    await page.fill('[aria-label="Password"]', 'SomePass');
    await page.click('[type="submit"]');

    await expect(page.getByText('Email is required')).toBeVisible();
  });

  // ── Acceptance: empty password → validation error ─────────
  test('empty password shows Password is required error', async ({ page }) => {
    await page.goto('/login');
    await page.fill('[aria-label="Email"]', 'user@example.com');
    await page.click('[type="submit"]');

    await expect(page.getByText('Password is required')).toBeVisible();
  });
});
```

Running git commit...
```bash
$ git add tests/login.spec.ts
$ git commit -m "feat: add login Playwright tests from QE runbook"
[main a3f21c9] feat: add login Playwright tests from QE runbook
 1 file changed, 52 insertions(+)
 create mode 100644 tests/login.spec.ts
✅ Tests committed successfully.
```

I'll verify the workspace state and provide a summary.

**Step 1 — git status**
```bash
$ git status
On branch main
nothing to commit, working tree clean
```

**Step 2 — git log**
```bash
$ git log -n 1
commit 341d284eb2b1f6ba4f5c7893840b81b57e9965ae
Author: Your Name <your.email@example.com>
Date:   Tue Jun 30 21:31:41 2026 +0100

    feat: add login Playwright tests from QE runbook
```

**Step 3 — Summary**
- **Automated**: Playwright E2E tests for the login feature in `./tests/login.spec.ts`.
- **Created Tests**: 5 test cases mapping to all acceptance criteria (success redirect, invalid password, empty email, empty password, session cookie verification).
- **Assumptions**: Assumed inputs have ARIA labels matching the login form fields (`Email`, `Password`).

✅ Verification complete. Workspace is clean and local history is recorded.

The QE has successfully directed Gemini 3.5 Flash through the full workflow: environment setup → workspace initialization → structured task definition → automated execution and commit.

The agent operated entirely under your observation and was correctable at every step via the markdown file. This is the QE as AI Orchestrator model in action.