Files
avdlee__swiftui-agent-skill/.github/copilot-instructions.md
T
2026-01-23 15:29:07 +01:00

4.4 KiB
Raw Permalink Blame History

GitHub Copilot Instructions: Agent Skills Expert (SwiftUI)

Review changes as an Agent Skills expert, focusing on the Agent Skills open format specification and ensuring this repository stays a SwiftUI-focused, non-architectural, modern-API skill.

Core Agent Skills Principles

Format Compliance

  • SKILL.md required: Every skill directory must contain a SKILL.md with YAML frontmatter.
  • YAML frontmatter requirements:
    • name: non-empty, <= 64 chars, lowercase letters/numbers/hyphens only.
    • description: non-empty, <= 1024 chars, clearly states what the skill does and when to use it.
  • Progressive disclosure: keep SKILL.md as the primary entrypoint; move deep dives into swiftui-expert-skill/references/.
  • Filesystem-based design: content should be organized so agents can load only what they need (metadata → instructions → references).

Skill Content Organization (Progressive Disclosure)

Level 1: Metadata (always loaded)

  • YAML frontmatter should be concise and trigger-relevant.

Level 2: Instructions (loaded when triggered)

  • swiftui-expert-skill/SKILL.md body should contain workflows, checklists, decision trees, and quick references.
  • Avoid turning SKILL.md into an encyclopedia; prefer linking out to references.

Level 3: Resources (loaded as needed)

  • Topic-specific guidance lives in swiftui-expert-skill/references/*.md.
  • References should be linked from SKILL.md so agents discover them naturally.

SwiftUI Repository Constraints (Keep This Skill “On Mission”)

  • SwiftUI-only: avoid UIKit patterns unless bridging is strictly necessary, and avoid server-side Swift.
  • No architecture mandates: do not prescribe MVVM/MVC/VIPER/coordinators/routers or project folder structures.
  • No tooling instructions: avoid IDE/tool walkthroughs. (Mentioning that profiling exists is fine; don’t give step-by-step tool instructions.)
  • Modern SwiftUI APIs: prefer modern, non-deprecated APIs where correctness is impacted (e.g., NavigationStack, foregroundStyle()).
  • Performance guidance is optional: performance optimizations should be phrased as “consider” unless it’s a correctness issue.

Review Focus Areas

  1. Description Quality

    • Does the description say what the skill does and when to use it (trigger conditions)?
    • Is it specific enough to be selected over other skills?
  2. Instruction Clarity

    • Are instructions actionable (checklists, decision trees, “review/improve/implement” flows)?
    • Are examples correct SwiftUI and consistent with the stated constraints?
  3. Progressive Disclosure

    • Is SKILL.md focused (not duplicating large reference content)?
    • Are deeper topics moved into references/ and linked appropriately?
  4. SwiftUI Correctness and Modernity

    • Modern API replacements are accurate and not misleading.
    • Guidance around lists/identity, sheets/navigation, and state wrappers is correct.
    • iOS 26+ / Liquid Glass guidance includes availability gating and sensible fallbacks.
  5. Safety / Security (for automation & scripts)

    • GitHub Actions and scripts should avoid surprising network behavior and avoid writing outside the repo.
    • Dependencies should be minimal and justified.

Common Issues to Flag

  • Missing/invalid YAML frontmatter in SKILL.md.
  • Invalid name (uppercase, underscores, spaces, too long).
  • Vague description (“helps with SwiftUI”) without clear trigger conditions.
  • Overlong SKILL.md that repeats reference content instead of linking to it.
  • Unlinked references in swiftui-expert-skill/references/ that aren’t discoverable from SKILL.md.
  • Architecture prescriptions or project-structure mandates.
  • Tooling walkthroughs (IDE-specific or step-by-step tool usage).
  • Non-SwiftUI deep dives (general Swift language/concurrency tutorials unrelated to SwiftUI usage).

Resources


Remember: You are reviewing as an Agent Skills expert. Prioritize format compliance, progressive disclosure, SwiftUI accuracy, and keeping guidance modern and non-prescriptive.