Back to directory

iOS Simulator Skill for Claude Code

A production skill of 29 scripts for building, testing, and automating iOS apps with Claude Code, wrapping Xcode builds and simulator interaction to save the token and context cost of raw xcodebuild output.

iOS Simulator Skill for Claude Code

iOS Simulator Skill for Claude Code covers both sides of iOS development: Xcode builds through xcodebuild, compiling, testing, and parsing results with progressive error disclosure so an agent does not have to read raw build output, and simulator interaction through xcrun simctl and idb, covering semantic UI navigation, accessibility testing, and device lifecycle management.

It installs through the Claude Code plugin marketplace or manually, either from a packaged release or from a git clone, and the skill itself lives one directory level inside the cloned repository, since the repository is a plugin containing the skill rather than the skill's root. Requirements include macOS 15 or later, Xcode with command-line tools, Python 3.12 or later for every script, and idb (CLI and companion) for tapping, swiping, and typing scripts, with Pillow needed only for visual diffs.

A companion MCP server version of just the Xcode build tooling, without the simulator scripts, is available separately as xclaude-plugin.

More in Agent skills

See all 581
  • Provenance gate

    github.com

    A deterministic authorization gateway that sits between an AI agent and its tools, tracing where each parameter of a side-effect call came from, and blocking the call or asking a human when the source does not check out.

  • A markdown framework of system rules that restates Isaac Asimov's Four Laws of Robotics as a set of prioritized loss calculations for large-language-model agents to follow, published as a discovery file for agents and crawlers to read.

  • Vitruvius

    github.com

    An engineering research agent that runs a discover, read, synthesize, verify and review loop over technical questions, reading sources directly rather than inferring from titles, and publishing its own scored verifier benchmark, weaknesses included.