All products

Draft QA test scenarios with AI

Give testers a repeatable starting state, explicit actions and the result to check at each step of a user story.

Draft QA test scenarios with AI from the behaviour the team has agreed on. An address change before dispatch makes a good example. The tester needs an editable order, an account allowed to change it and somewhere to verify the saved address afterwards. The test-scenarios skill turns those details into instructions a colleague can repeat.

A success message answers only one question. If persistence matters, include reopening the order in the procedure. If the specification says nothing about an order already in transit, record the missing decision before treating an imagined error message as a requirement.

What it does

  • Define the objective of each scenario.
  • Specify the starting state and the role performing the action.
  • Pair each action with an outcome the tester can observe.
  • Cover invalid input and boundary conditions separately.
  • Trace the scenario back to the acceptance criterion it checks.

Install and use

  1. Install the skill in your agent. Give your agent the product context, user story and acceptance criteria. Drafting requires no separate testing service.
  2. Ask for one editable order and one dispatched order, with explicit prerequisites for each.
  3. Have a colleague run the cases with disposable test data and record actual results separately from expectations.

FAQ

Source

Made by Paweł Huryn. Original skill: test-scenarios, published on GitHub under the MIT licence.

View on GitHub