swiftui-audit

name	swiftui-audit
description	Run the full evidence-backed macOS SwiftUI audit suite over a project. Use only when explicitly requested with $swiftui-audit.

Full macOS SwiftUI audit

Bundled resource root

Let <swiftui-plugin-root> be the absolute plugin directory two levels above this SKILL.md. Resolve that path before running commands and shared references.

1. Resolve the requested target, defaulting to the current directory. Create the output directory and persist the relevance scan:

   mkdir -p swiftui-audits
   python3 "<swiftui-plugin-root>/scripts/audit-scan.py" "<target>" --json swiftui-audits/_scan.json
   

2. Read <swiftui-plugin-root>/skills/audit-macos-swiftui-full/SKILL.md completely. Follow that orchestrator using _scan.json and preserve its dependency waves: guards; state and data; UI domains; boundaries and scoring.

3. When collaboration tools are available, run independent auditors in the same wave concurrently. Otherwise run them serially in the listed order.

4. Write per-finding Markdown under swiftui-audits/ using <swiftui-plugin-root>/references/_shared/finding-schema.md, then create _SUMMARY.md and its 0–100 nativeness score.

5. Read .swiftui-plugin/settings.md, falling back to .claude/swiftui.local.md. With strict_audit: true or no setting, label hard findings release-blocking. With strict_audit: false, label the same findings advisory. The mechanical audit-gate.sh accepts matching --strict and --advisory modes for CI exit behavior.

6. Apply fixes only when the user requested them and the shared fix-safety protocol permits them. Keep all other findings open.

The lint engine locates candidates; every reported finding still requires source inspection and verification through the bundled corpus and current Apple documentation.