{
  "id": "page:docs-reference-repos-twostraws-swiftui-agent-skill-research",
  "_kind": "Page",
  "_file": "wiki/docs/reference-repos/twostraws/swiftui-agent-skill/research.md",
  "_cluster": "wiki",
  "attributes": {
    "nodeKind": "Page",
    "sourcePath": "docs/reference-repos/twostraws/SwiftUI-Agent-Skill/research.md",
    "sourceKind": "repo-docs",
    "title": "twostraws/SwiftUI-Agent-Skill",
    "displayName": "twostraws/SwiftUI-Agent-Skill",
    "slug": "docs/reference-repos/twostraws/swiftui-agent-skill/research",
    "articlePath": "wiki/docs/reference-repos/twostraws/SwiftUI-Agent-Skill/research.md",
    "article": "\n# twostraws/SwiftUI-Agent-Skill\n\n- **Archetype**: domain-skill-pack\n- **Stars**: 3,494\n- **Last pushed**: 2026-03-11\n- **License**: MIT\n- **Discovered**: 2026-04-12\n- **Skills found**: 1 SKILL.md (swiftui-pro)\n- **Fork**: No\n- **Source**: gh-search\n\n## Summary\n\nA single focused SwiftUI agent skill by Paul Hudson (Hacking with Swift). Targets common LLM mistakes when generating SwiftUI code: deprecated API usage, VoiceOver accessibility issues, performance problems. Covers navigation, layout, animations, state management, accessibility, and deprecated API detection. Part of a larger Swift Agent Skills ecosystem (SwiftData, Swift Concurrency, Swift Testing).\n\nUses the Agent Skills standard. Installable via `npx skills add`.\n\n## Assessment\n\nLow-to-moderate extractable value for babysitter. Very domain-specific (iOS/SwiftUI only). However, the pattern of \"targeting mistakes LLMs actually make\" is valuable procedural knowledge. The skill demonstrates a quality-gate approach focused on LLM-specific failure modes rather than general best practices.\n\n**Extraction priority**: LOW\n\n---\n\n## Processes\n\n### 1. SwiftUI Code Quality Review\n- **Source skills**: swiftui-pro\n- **Placement**: `specializations/shared/swiftui-code-quality-review.js`\n- **Description**: Quality gate process for SwiftUI projects focused on LLM-generated code: deprecated API detection -> accessibility/VoiceOver audit -> performance review (state management, layout) -> navigation pattern check -> animation safety review. Generates structured findings with severity levels.\n\n## Library Mapping\n\n| Extractable Process | Library Status | Action | Existing Path | Target Placement |\n|-------------------|----------------|--------|---------------|------------------|\n| SwiftUI Code Quality Review | NEW | Quality gate process for SwiftUI focused on LLM-specific mistakes | - | specializations/mobile/swiftui-code-quality-review.js |\n| LLM-Mistake-Targeted Review | NEW | Review framework focused on what AI gets wrong rather than general best practices | - | specializations/shared/llm-mistake-targeted-review.js |\n\n## Plugin Marketplace Mapping\n\n| Plugin Idea | Marketplace Status | Action | Existing Plugin | Target Placement |\n|-------------|-------------------|--------|-----------------|------------------|\n| N/A | N/A | No valid plugin ideas - purely instructional content with no tool integration surface | - | N/A |\n\n## Implicit Procedural Knowledge\n\n- **LLM mistake targeting**: The skill is explicitly designed around \"mistakes LLMs actually make\" rather than general best practices. This targeted-correction pattern is valuable for designing quality gate processes in any domain -- focus on what the AI gets wrong, not everything.\n- **Ecosystem skill families**: Part of a coordinated set (SwiftUI, SwiftData, Swift Concurrency, Swift Testing) showing how domain skills can be decomposed into focused, composable units. Relevant for how babysitter organizes domain specializations.\n",
    "documents": []
  },
  "outgoingEdges": [],
  "incomingEdges": [
    {
      "from": "page:docs-reference-repos",
      "to": "page:docs-reference-repos-twostraws-swiftui-agent-skill-research",
      "kind": "contains_page"
    }
  ]
}