Turns test scenarios into actual tasks people can complete without being led to the answer. You get context setting that motivates without hinting, success criteria you can measure (time, errors, completion rate), and observation guides for what to watch. It pushes you away from asking questions and toward action-oriented goals, which is the difference between "How would you find the return policy?" and "Find the return policy." Includes task types from exploratory to open-ended and rules like no product jargon, one goal per task, and ordering easy to hard. Honestly most useful if you're running usability tests and keep accidentally writing tasks that telegraph the UI solution.
npx skills add https://github.com/owl-listener/designer-skills --skill test-scenario