What problem does it solve? Automated tests can all pass while a feature is visibly broken in the browser—missing toolbox entries, wrong colors, misplaced buttons, or stale localStorage state. This Skill closes that gap by performing structured in-browser acceptance checks before declaring work done. ## Core Features & Use Cases - Programmatic enumeration: Queries the live application (e.g., toolbox contents via the app's own assembly functions) instead of eyeballing counts, avoiding toggle and scroll-viewport traps. - Mandatory visual screenshots: Captures every toolbox category, a dragged-out block, and cross-language comparisons, since layout and mutation-button issues are only detectable by eye. - Execution-path verification: When changes affect runtime behavior (pause, input, fallback, error handling), it checks four extra things: where execution stops, whether variable state is visible, whether edits take effect on continue, and whether manual intervention leaves a trace. - Use Case: After adding new Blockly blocks to a course, run this Skill to confirm the blocks actually appear in the toolbox, render correctly when dragged out, and round-trip through code conversion without loss. ## Quick Start Verify the feature I just finished by opening it in the browser, clearing localStorage, enumerating the toolbox programmatically, and screenshotting each category.
