Manual test case, markdown. Test cases are plain markdown files. Edit them like any document, review changes in git, sync with Testomat.io.

<!-- suite -->
# Checkout

<!-- test
type: manual
priority: high
-->
# Guest can pay with a saved card

## Preconditions

- The cart has one product.

## Steps

- Open the cart and click Checkout
  *Expected*: The payment form is shown
- Pick the saved card and click Pay
  *Expected*: The order confirmation page opens

Automated test. Automated tests live next to the manual ones. The agent can write them from manual cases and fix them when they fail.

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

test('guest can pay with a saved card', async ({ page }) => {
  await page.goto('/cart');
  await page.getByRole('button', { name: 'Checkout' }).click();
  await page.getByLabel('Saved card').check();
  await page.getByRole('button', { name: 'Pay' }).click();
  await expect(page).toHaveURL(/order-confirmed/);
});

Application code. With the repository in the workspace the agent reads the code behind a feature, so its test ideas match what was built.

export function cartTotal(items: CartItem[], coupon?: Coupon) {
  const sum = items.reduce((t, i) => t + i.price * i.qty, 0);
  if (!coupon) return sum;
  if (coupon.expiresAt < Date.now()) return sum;
  return Math.max(0, sum - coupon.amount);
}

Local document. Add specifications and requirements as extra context. The agent uses them to understand the task.

# Checkout requirements

- A guest can pay without creating an account.
- Saved cards are offered first.
- An expired coupon is ignored and the user is told why.
- The order confirmation shows the order number.

The full test inventory with suite, priority and tag filters.

Run history. Open a run to see results, or continue a manual run step by step.

Launch a run from a plan without leaving the app.