What problem does it solve? Diagnosing why a QML / Qt Quick UI feels laggy, drops frames, or stutters is hard without tooling: this Skill runs qmlprofiler end-to-end, parses the resulting .qtd trace, and maps the top hotspots back to your project source with concrete fixes. ## Core Features & Use Cases - End-to-end profiling workflow: Locates the Qt toolchain, builds with -DQT_QML_DEBUG, runs qmlprofiler with selectable profiles (full, rendering, logic, memory), and falls back to a manual command when the session cannot execute binaries. - Trace parsing and summaries: A bundled Python parser emits JSON covering event-type totals, animation frame-time percentiles (p50/p95/p99), memory allocation by QV4 category, and pixmap cache usage. - Hotspot analysis and reporting: Maps the top 30 hotspots to local QML files, analyzes the top 5 against an anti-pattern catalogue (Binding, Javascript, HandlingSignal, Creating, Compiling, SceneGraph/Painting, Memory/Pixmap), and writes a timestamped standalone Markdown report. - Use Case: A developer notices their Qt Quick app stutters during scrolling; they invoke the Skill, exercise the app, and receive a report showing a Binding hotspot in a delegate with a suggested fix. ## Quick Start Ask your assistant to profile your QML application by passing the executable (for example, profile my app with the qt-qml-profiler using -- build/myapp) or supply an existing .qtd trace file for analysis only.
