SwiftUI Expert Agent Skill
Actionable SwiftUI best-practice guidance for AI coding tools
Covers SwiftUI review, modern APIs, accessibility, performance and Instruments trace analysis.
Rules are split by topic and trace scripts are inspectable; fixed-revision scripts passed syntax checks, with no automated quality suite.
Core review is read-only with no active network use; performance recording invokes xctrace, launches or attaches to a process and writes trace files.
Implementation and references are transparent, but workflows focus on release and sync rather than outcome benchmarks.
A concise entry point, progressive references and clear script arguments work well when macOS and Xcode are available.
Recently maintained for new APIs, with concentrated ownership and limited regression evidence.
- SwiftUI APIs change quickly; verify against Apple documentation for the deployment target and compile the result.
- Confirm device, process and output path before recording; traces may contain sensitive runtime details.
- Performance scripts require macOS with Xcode Instruments.
What it does & when to use it
SwiftUI Expert Skill provides guidance for AI coding tools that support the Agent Skills open format. It covers SwiftUI state management, view composition, performance, navigation, animation, accessibility, image optimization, and iOS 26+ Liquid Glass adoption. The repository also includes Python and xctrace tooling for recording and analyzing Instruments traces. It is explicitly non-opinionated about architecture and code style, focusing instead on correctness and performance.
The skill organizes SwiftUI best practices into an SKILL.md workflow and on-demand reference files for code review, refactoring, and troubleshooting. Topics include state management, view structure, lists, scrolling, sheets and navigation, Swift Charts, animations, macOS scenes, Liquid Glass, accessibility, and image optimization. Its bundled trace tooling can record and analyze Time Profiler, Hangs, Animation Hitches, SwiftUI update, and cause-graph data from Instruments traces.
- Review SwiftUI state-management and data-flow choices.
- Refactor view hierarchies for better composition and stable identity.
- Investigate performance issues involving lists, scrolling, animations, or SwiftUI updates.
- Migrate deprecated SwiftUI APIs and evaluate iOS 26+ Liquid Glass patterns.
- Review Swift Charts, navigation, sheets, macOS windows, and multi-window scenes.
- Record or analyze `.trace` files to investigate hangs, hitches, and expensive SwiftUI view updates.
Pros & cons
- Broad coverage spanning state management, view composition, Charts, macOS, accessibility, and Liquid Glass.
- Reference files are designed to load on demand, which the README presents as a way to limit unnecessary task context.
- Includes xctrace recording, parsing, cross-lane correlation, and Markdown summary tooling for SwiftUI performance investigations.
- Documents installation paths for skills.sh, Claude Code, Cursor, Codex, pi, and manual setups.
- MIT licensed and accompanied by a workflow for maintaining API guidance.
- SKILL.md was not cached in the supplied material, so its complete frontmatter, instructions, and exact workflow cannot be independently verified here.
- Trace analysis requires Python 3 and xctrace from Xcode, making that feature dependent on an Apple development environment.
- The maintenance workflow requires the Sosumi MCP; this is not stated as a requirement for every normal skill-use scenario.
- The README does not provide a verified version compatibility matrix; the `weekly installs 16.6k` badge is not evidence of feature coverage or quality.
How to install
With skills.sh, run npx skills add https://github.com/avdlee/swiftui-agent-skill --skill swiftui-expert-skill. In Claude Code, run /plugin marketplace add AvdLee/SwiftUI-Agent-Skill, then /plugin install swiftui-expert@swiftui-expert-skill. For Codex or compatible OpenAI tools, copy or symlink swiftui-expert-skill/ into the skills directory, for example cp -R swiftui-expert-skill/ "$CODEX_HOME/skills/swiftui-expert-skill". The repository also supports pi install https://github.com/AvdLee/SwiftUI-Agent-Skill, or manual installation after cloning the repository.
How to use
After installation, ask the agent to use the swiftui expert skill, for example: Use the swiftui expert skill and review the current SwiftUI code for state-management and performance improvements. The agent should use the workflow and checklists in swiftui-expert-skill/SKILL.md and load the relevant reference file for the task. For trace work, provide a .trace path; the recording and analysis scripts are scripts/record_trace.py and scripts/analyze_trace.py.
Compared to similar skills
Compared with generic SwiftUI code-generation prompts, this repository provides a structured Agent Skill, topic-specific reference files, and performance trace tooling. It is explicitly non-opinionated about architecture and code style, emphasizing correctness and performance instead. Compared with text-only guidance, it adds xctrace recording and analysis capabilities, although those tools require Xcode.