Test Cases

Manual test case creation

Write precise instructions for a browser-executed or manually tracked workflow.

Manual creation is the best option when you already know the workflow and expected result.

Write the steps

Write each step with an Action and an Expected result. For AI-run tests, the action can describe user intent instead of low-level clicks:

1. Action: Add the first product in the catalog to the cart.
   Expected result: The cart shows the first product as added.
2. Action: Open the cart.
   Expected result: The cart contains the selected product with quantity 1.

The action should describe what the AI should accomplish. The expected result should describe the visible evidence that proves the step succeeded. Avoid relying on animation timing, random data, or hidden system state.

Use credentials

Select a credential when the workflow begins behind login or requires a role. Credential-specific login instructions are combined with the test case instructions during execution.

Choose run type

Use AI for checks Killbug should execute in a browser. Use Manual when the test case documents coverage that a human or external process executes.

On this page