[ Manual testing ]
Manual testing
[Intro: test tools, resources, and widget responses without writing any test code.]
Tool calls
[How to pick a tool from the sidebar, fill in arguments, and run it. Mention the Args / Definition tabs.]
Resources
[How to pick a resource, enter a URI, and read its contents. Mention the result pane.]
Widget resource
[Explain widget resources: inject mock tool data, pick platform (Claude/OpenAI), theme, viewport, and interact live in the preview pane.]
Test data tab
[Describe the Test Data editor: inject toolInput, toolOutput, and widgetState as JSON to simulate different AI responses without calling a real model.]
Platform and viewport
[Describe platform selector (Claude / OpenAI) and viewport presets (Mobile / Tablet / Desktop / Custom). Mention compatibility badges.]