What problem does it solve? macOS SwiftUI code written by AI often compiles and looks plausible but reads like an iPad app dropped into a window, missing pointer, menu, window, and navigation affordances that define a native Mac experience. This Skill measures that gap with a deterministic 0-100 nativeness score and routes every finding to the owner skill that fixes it. ## Core Features & Use Cases - 15-smell detection catalog: Detects absent Mac affordances such as missing .onHover, .contextMenu, .help tooltips, .formStyle(.grouped), Table, Settings scenes, and .commands menus across five weighted categories. - Deterministic 0-100 scoring dashboard: Computes a nativeness score with per-category deductions and writes a prioritized punch-list as a nativeness-dashboard index file. - Route-not-fix discipline: Every finding carries a cross_ref to the owner audit skill (pointer-gestures, controls-forms, layout-and-tables, navigation-toolbars, menus-commands, scenes-windows) that owns the actual fix. - Use Case: Point it at a finished macOS SwiftUI project to learn that it scores 68/100 because it uses a NavigationStack as its shell, lacks a Settings scene, and has rows with no hover or right-click affordances, then run the routed owner skills to raise the score. ## Quick Start Audit my macOS SwiftUI project for iPad-in-a-window smells and give me the nativeness score with a prioritized punch-list.
