{
  "markdown": "# iOS Performance Agent Skills\n\n<p align=\"center\">\n<img width=\"320\" src=\"https://github.com/user-attachments/assets/a71688d2-54a3-4ea5-8362-e4659b3d46c6\"/>\n</p>\n\nA collection of AI-agent skills for reviewing, diagnosing, and improving performance in iOS applications.\n\nEach skill lives in its own top-level directory and installs with a single `npx` command (see [Installation](#installation)), or can be copied manually into another agent setup.\n\n## What this is\n\nThis repository contains practical performance-focused instructions for AI coding agents.\n\nThe skills are designed to help an agent reason about iOS performance problems with a more structured mental model: what to inspect, what to avoid assuming, which trade-offs matter, and when measurement is required.\n\nIt is a collection of reusable guidance files for AI-assisted code review, refactoring, and investigation.\n\n## Included skills\n\n### `ios-launch-performance`\n\nFor analyzing app startup and launch-time regressions.\n\nCovers cold, warm, and prewarmed launch, dyld and pre-main work, static initializers, AppDelegate, SceneDelegate, SwiftUI app entry points, first frame, SDK initialization, launch metrics, and validation with tools such as Instruments, XCTest, and MetricKit.\n\n### `ios-performance-profiling`\n\nFor choosing and interpreting performance profiling workflows.\n\nCovers Instruments, traces, XCTest metrics, MetricKit payloads, signposts, memory graphs, animation hitches, hangs, CPU work, allocations, disk I/O, networking, and production performance signals.\n\n### `ios-perceived-performance`\n\nFor improving how fast and responsive an app feels to users.\n\nCovers loading states, first feedback, progressive rendering, skeletons, optimistic updates, high-stakes actions, perceived latency, and validation through recordings, traces, and user-visible behavior.\n\n### `swiftui-performance`\n\nFor reviewing SwiftUI code composition and update behavior.\n\nCovers identity, state ownership, dependency scope, body cost, list performance, row complexity, bindings, closures, layout, drawing, animations, async lifecycle work, and profiling validation.\n\n### `swift-concurrency-performance`\n\nFor reviewing performance and correctness trade-offs in Swift Concurrency.\n\nCovers actor isolation, MainActor usage, task lifecycle, task groups, cancellation, AsyncSequence, continuations, reentrancy, executor behavior, and concurrency-related responsiveness issues.\n\n### `swift-runtime-performance`\n\nFor reviewing Swift runtime-level performance costs.\n\nCovers allocations, ARC traffic, stack vs heap behavior, dispatch, existentials, generics, opaque types, copy-on-write, SIL optimization, unsafe Swift, modularization, linking, and launch-time trade-offs.\n\n## Installation\n\nInstall the whole bundle with one `npx` command:\n\n```bash\nnpx skills add Livsy90/iOS-Performance-Agent-Skills --all\n```\n\nOr install only the skills you want:\n\n```bash\nnpx skills add Livsy90/iOS-Performance-Agent-Skills --skill ios-launch-performance --skill swiftui-performance\n```\n\nThe CLI will ask which agents to register the skill with (Claude Code, Codex, Cursor, Gemini, ...) and whether to install per-project or globally.\n\nIf `npx` is missing, install Node (`brew install node`); if `brew` is missing too, [install Homebrew](https://brew.sh) first.\n\n### Alternative install methods\n\n**Claude Code** (installs the whole bundle directly):\n\n```bash\n/plugin install Livsy90/iOS-Performance-Agent-Skills\n```\n\nOr clone this repository and drop the skill folders wherever your agent expects them.\n\n## Repository structure\n\n```text\nios-launch-performance/\n  SKILL.md\n  references/\n\nios-performance-profiling/\n  SKILL.md\n  references/\n\nios-perceived-performance/\n  SKILL.md\n  references/\n\nswiftui-performance/\n  SKILL.md\n  references/\n\nswift-concurrency-performance/\n  SKILL.md\n  references/\n\nswift-runtime-performance/\n  SKILL.md\n  references/\n\n.claude-plugin/plugin.json    # Claude Code plugin manifest\ngemini-extension.json         # Gemini extension manifest\n```\n",
  "bytes": 4007,
  "sha": "6d95aab1f3a666ae000469f2394a1d2bd069dd3e4b3dc08e0934b41e6733e66f",
  "repo_slug": "livsy90/ios-performance-agent-skills",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/plg_livsy90_ios_performance_agent_skills_6f8d2760/readme"
}