This is a systematic framework for running usability tests with real users watching them attempt core tasks. You get test planning templates, think-aloud protocol instructions, task scenario examples that avoid hand-holding, and a severity rating system based on impact times frequency. The approach is grounded in Nielsen's research about diminishing returns after five users. What I appreciate is the clarity on what makes a good versus bad task scenario. Writing "find all conversations with Acme Corp" tests discoverability, while "click Settings then click Invite" just makes people follow orders. Includes the full System Usability Scale questionnaire and a report template tracking success rates, time on task, and error counts.
npx skills add https://github.com/daffy0208/ai-dev-standards --skill usability-tester