By wei18
30 first-party Claude Code skills for AI-agent-driven Swift / Apple-platform development — Swift 6 / SwiftPM / testing / CI / L10n / telemetry defaults, Leader-Developer collaboration patterns, plus accessibility, dependency injection, and performance engineering. Distilled from a real shipping project and public Apple/standards docs.
Maintain a running `meetings/{date}_{topic}.impl-notes.md` *during* subagent task execution to capture in-flight design decisions, intentional deviations from spec, considered alternatives, and open questions for Leader/User. Distinct from the post-hoc phase meeting log; this file is updated incrementally as decisions are made. Invoke at the start of any non-trivial subagent dispatch, when ambiguity is encountered mid-task, or when about to deviate from spec.
Default localization scope AND execution playbook for Apple-platform Apps — 7 locales (en, zh-Hant, zh-Hans, ja, ko, es, th) translated via AI agent flow using `Localizable.xcstrings`. Source = en, primary = zh-Hant. Minimum set zh-Hant + en. Invoke when (1) deciding L10n scope / catalog format / translation flow at project setup, OR (2) actually executing a translation pass to add or refresh strings.
Rasterize a designer-authored 1024×1024 SVG into the PNG that Apple's asset catalog wants — no Homebrew dependency, no Cloud service, just `qlmanage` already on the Mac.
Use when preparing an App Store submission, diagnosing an App Review rejection / Resolution Center message, or hardening a free game/app with ads + Remove-Ads IAP + CloudKit + Game Center against the rejection classes it realistically hits. Covers guideline 1.3 / 2.1 / 2.3 / 2.5.x / 3.1.1 / 4.0 / 4.3 / 5.1.1 / 5.1.2, the ATT-vs-AdMob trap, privacy-label parity, and a pre-submit checklist.
Default minimum deployment targets for Apple-platform Swift Apps and how to deviate up or down. Default is iOS 18 / macOS 15 with Xcode 16+ (first toolchain to formally support Swift 6 language mode). Invoke when starting a new Apple-platform project, writing Package.swift `platforms:` list, deciding whether to adopt Liquid Glass / latest-OS-only APIs, or when asked "what should the minimum iOS / macOS version be?".
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
A curated catalog of Swift / Apple-platform skills for AI coding agents (Claude Code). It ships original first-party skills for one project's hard-won patterns and for public-standard topics, plus aggregated best-of-breed external skill plugins referenced from their original repos and credited to their authors — not copied.
繁體中文版:
README.zh-Hant.md
The repo is two things at once:
apple-dev-skills) — 30 first-party SKILL.md files under
skills/, namespaced apple-dev-skills:<skill>. These were distilled from a real
shipping project's own experience and from public Apple / WCAG / Swift documentation.This README is the single source of truth (SSOT) for the library.
| Phase | What | Status |
|---|---|---|
| 1 — Extract | Pull one project's portable skills out, genericize, ship as a submodule-consumable plugin | ✅ done |
| 2 — Standalone library | README as SSOT · an npx skills add install path · a marketplace catalog that can aggregate other repos (mechanism in claude-skill-plugin-packaging) | ✅ done |
| 3 — Curate & aggregate | Surveyed high-star Swift/Apple skill repos and aggregated the best-of-breed MIT plugins by reference (credited to their authors — see below), rather than re-implementing them. First-party skills were written only for genuine gaps with no aggregatable plugin (accessibility, dependency injection, performance). Decisions: CURATION.md | ✅ done |
Detailed status, spec→impl alignment, and future phases: ROADMAP.md.
Claude Code discovers shared skills through the plugin system, not by a bare folder under
.claude/skills/. This repo is both a plugin and a marketplace.
/plugin marketplace add wei18/apple-dev-skills
/plugin install apple-dev-skills@apple-dev-skills
Our skills load as apple-dev-skills:<skill-name>. The same marketplace also offers
the aggregated external plugins — install any of them
from their authors' repos, e.g. /plugin install swiftui-expert@apple-dev-skills.
Vendor the first-party plugin into one repo and pin its version, then register it as a local-path marketplace at project scope so it loads from the pinned submodule:
git submodule add https://github.com/wei18/apple-dev-skills.git .claude/skills/apple-dev-skills
cd .claude/skills/apple-dev-skills && git checkout v0.5.0 && cd -
Then commit this to the repo's .claude/settings.json (collaborators are prompted
to trust the workspace, after which it auto-registers):
{
"extraKnownMarketplaces": {
"apple-dev-skills": {
"source": { "source": "directory", "path": "./.claude/skills/apple-dev-skills" }
}
},
"enabledPlugins": { "apple-dev-skills@apple-dev-skills": true }
}
A bare git submodule add alone will not surface the skills — the marketplace
registration above is what makes Claude Code load them.
npx skills (flat skills, no plugin)Copy chosen first-party SKILL.md files into your agent's skills directory with the
open skills CLI, flat and un-namespaced:
npx skills add wei18/apple-dev-skills --list # browse the full index
npx skills add wei18/apple-dev-skills --skill swift6-concurrency # one skill
npx skills add wei18/apple-dev-skills --all -a claude-code # everything
npx claudepluginhub wei18/apple-dev-skills --plugin apple-dev-skillsRun the Upkeep repo audit locally against any repository path — output-only drift report with evidence and severity.
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Develop, test, build, and deploy Godot 4.x games with Claude Code. Includes GdUnit4 testing, web/desktop exports, CI/CD pipelines, and deployment to Vercel/GitHub Pages/itch.io.
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review
Lazy senior dev mode. Forces the simplest, shortest solution that actually works: YAGNI, stdlib first, no unrequested abstractions.
Upstash Context7 MCP server for up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your LLM context.
A growing collection of Claude-compatible academic workflow bundles. Covers scientific figures, manuscript writing and polishing, reviewer assessment, citation retrieval, data availability, paper reading, literature search, response letters, paper-to-PPTX conversion, and evidence-grounded Chinese invention patent drafting. Rules are organized as reusable skill folders with explicit workflows and quality checks.