qt development skills

Qt AI Skills

Official agentic skills for Qt software development and quality assurance, designed for use with AI coding tools such as Claude Code, Codex CLI, and GitHub Copilot.

Skills have been tested with frontier LLMs from the Claude, Gemini, and GPT model families.

There is no settled industry standard for AI skill packaging. Each platform has its own conventions. Our canonical format uses SKILL.md with YAML frontmatter in a directory-based structure — this works natively on Claude Code, Codex CLI, and GitHub Copilot, and can be adapted to other platforms through condensed variants. See CONTRIBUTING.md for the full cross-platform story.

These agentic development skills use AI and can make mistakes. Always double-check the output carefully. Before using the skills under Qt commercial licensing, make sure you have understood and agree to the Qt AI Services Terms & Conditions. By using the skills or MCP tools, you accept these terms & conditions and that you have the right to do so on behalf of your employer.

Skills

| Skill | Type | Description | |-------|------|-------------| | qt-cpp-review | Review | Deterministic linting + deep analysis for Qt C++ code (inline for small scopes, parallel shard agents above about 2,500 lines). Covers model rule compliance, memory ownership, thread safety, correctness, error handling, and performance. | | qt-qml-review | Review | Deterministic QML linting (48+ rules) + deep analysis (inline for small scopes, parallel shard agents above about 1,500 lines) for bindings, layout, loaders, delegates, states, and performance. | | qt-qml | Conceptual | QML best practices for writing, reviewing, fixing, and refactoring. Corrects systematic LLM pre-training biases around bindings, scoping, modules, JS interop, and types. | | qt-canvas2d | Conceptual | Qt Canvas2D (Qt 6.12+, Qt Canvas Painter) best practices for GPU-accelerated imperative 2D drawing in QML — gauges, charts, waveforms, freehand drawing. Corrects LLM confusion with the legacy Qt Quick Canvas and HTML5 <canvas> APIs | | qt-ui-design | Conceptual | UI design and audit for Qt/QML, web, and embedded (MPU/MCU) targets. Covers screen layout, navigation, and UX review with platform-aware defaults for geometry, viewing distance, input, and locale. | | qt-qml-docs | Process | Generates Markdown reference documentation for QML components and applications from .qml source files. | | qt-cpp-docs | Process | Generates Markdown reference documentation for Qt/C++ source files — classes, modules, utilities, headers, and entry points. | | qt-qml-profiler | Tool | Runs qmlprofiler on a 2D QML / Qt Quick application, parses the .qtd trace, and analyzes hotspots against the source code with frame-time, memory, and pixmap-cache summaries. Does not cover Qt Quick 3D. | | qt-qml-test | Process | Generates Qt Quick Test cases (tst_*.qml) for QML components using TestCase, SignalSpy, and tryCompare. Handles single files and batches. Does not cover CMake or runner setup. | | qt-qml-test-run | Tool | Builds and runs Qt Quick Test (qmltestrunner) tests for a CMake project, parses the JUnit XML, and writes a Markdown report. Opt-in CMake test-infrastructure wiring with --wire-up. Companion to qt-qml-test. | | qt-figma-token-extraction | Process | Extracts design tokens, text styles, and variables from a Figma design system and produces a design-tokens.json plus ready-to-use QML singletons. | | qt-figma-component-generation | Process | Extracts component metadata from a Figma design system and generates production-ready QML controls mapped to Qt Quick Controls 2 patterns. Requires tokens from qt-figma-token-extraction. | | qt-cmake-project | Conceptual | Sets up and manages Qt 6 projects built with CMake — fresh projects, executables, libraries, QML modules, plugins, folder layout, and static resources. Corrects systematic LLM biases around qmake-isms and the legacy qt5_* macros. |

Skill types

• Review — structured code review workflows combining

deterministic linters with deep AI analysis

• Process — workflows and decision frameworks

(architecture, build, test, documentation)

• Conceptual — mental model corrections for areas where

LLMs consistently fail (declarative QML, C++/QML boundary, Widgets patterns, UI design)