ios-simulator-skill: Enhance Claude's iOS App Development with Simulator Skill

Introduction

The ios-simulator-skill is a powerful, production-ready skill for Claude Code, designed to significantly enhance the ability of AI agents to build, test, and automate iOS applications. This repository offers a comprehensive suite of 27 scripts, optimized for both human developers and AI agents, to streamline iOS development workflows.

It covers both Xcode builds, via xcodebuild, for compiling, testing, and parsing results with progressive error disclosure, and Simulator interaction, via xcrun simctl and idb, for semantic UI navigation, accessibility testing, and device lifecycle management. A key benefit is its token-efficient output, which drastically reduces the amount of information LLMs need to process, saving on token wastage and focusing agent conversations.

Installation

Via Plugin Marketplace (Recommended)

In Claude Code, use the following commands:

/plugin marketplace add conorluddy/ios-simulator-skill
/plugin install ios-simulator-skill@conorluddy

Via Git Clone

For personal or project-specific installations:

# Personal installation
git clone https://github.com/conorluddy/ios-simulator-skill.git ~/.claude/skills/ios-simulator-skill

# Project installation
git clone https://github.com/conorluddy/ios-simulator-skill.git .claude/skills/ios-simulator-skill

After cloning, restart Claude Code for the skill to load automatically.

Prerequisites

Ensure your system meets these requirements:

macOS 12+
Xcode Command Line Tools (xcode-select --install)
Python 3
IDB (optional, for interactive features: brew tap facebook/fb && brew install idb-companion)
Pillow (optional, for visual diffs: pip3 install pillow)

Examples

The skill's operational limits are tunable via IOS_SIM_* environment variables, allowing adaptation to different environments, from local development to slow CI runners. Here are some examples:

Slow GitHub Actions macOS runner: Give boot up to 10 minutes.

IOS_SIM_BOOT_TIMEOUT=600 python scripts/simctl_boot.py --wait-ready

Monorepo with hundreds of warnings: See them all in verbose mode.

IOS_SIM_BUILD_VERBOSE_CAP=500 python scripts/build_and_test.py --verbose

Complex Settings-style screen: Return more tappable elements.

IOS_SIM_MAX_ELEMENTS=100 python scripts/navigator.py --list-tappable

Snappy app: Cut tap-settle delay in half for faster E2E runs.

IOS_SIM_TAP_SETTLE_MS=250 python scripts/navigator.py --find-text "Login" --tap

Long CI pipeline: Keep cache entries valid for the whole job.

IOS_SIM_CACHE_TTL_HOURS=8 python scripts/build_and_test.py --project MyApp.xcodeproj

You can also export these variables once for the entire session:

export IOS_SIM_BOOT_TIMEOUT=600
export IOS_SIM_LOG_TAIL=500
export IOS_SIM_MAX_ELEMENTS=50
# … all subsequent scripts honor them

Why Use It

The ios-simulator-skill offers several compelling advantages for developers and AI agents working with iOS applications:

Optimized for AI Agents: Specifically designed for Claude Code, it enables AI to interact with iOS simulators and Xcode builds efficiently, reducing the cognitive load and token cost for LLMs.
Token Efficiency: Features like progressive disclosure for build logs, optimized screenshot compression, and accessibility-first UI navigation drastically cut down on token consumption, making AI interactions more cost-effective and faster.
Robust UI Interaction: Instead of fragile pixel-coordinate tapping, the skill uses iOS accessibility APIs for semantic navigation, ensuring interactions are robust and resilient to UI changes. For more details on accessibility-first navigation for AI agents, refer to AI-Accessible Apps.
Comprehensive Tooling: With 27 specialized scripts, it provides extensive capabilities for build management, device state manipulation, app navigation, testing, privacy management, and simulator lifecycle control.
Configurable Environment: Nearly every operational limit is tunable via environment variables, allowing fine-grained control to adapt the skill's behavior to diverse development and CI/CD environments.
Proven Reliability: Tested using Claude Code evals, the skill demonstrates a 100% pass rate, ensuring its reliability and effectiveness in automated workflows.