Ios Device Automation

Automates iOS devices through natural language commands and screenshot analysis instead of traditional DOM scraping or accessibility APIs. You connect a device, Claude takes screenshots, then issues commands like "tap the login button" or "scroll to the bottom" based on what it sees. The workflow is strictly synchronous: one command at a time, wait for the screenshot, analyze, then act. It's built on Midscene and needs a vision model like Gemini or Qwen configured via environment variables. The critical thing to know is that each command can take up to a minute because of AI inference, so this isn't for speed. Good for testing apps where traditional automation hooks don't exist or when you want to drive interactions purely from visual state.

skills.sh installs: 3.2K
GitHub stars: 321
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add web-infra-dev/midscene-skills --skill ios-device-automation --agent claude-code

Installs into .claude/skills of the current project.

Files

iOS Device Automation

CRITICAL RULES — VIOLATIONS WILL BREAK THE WORKFLOW:

Never run midscene commands in the background. Each command must run synchronously so you can read its output (especially screenshots) before deciding the next action. Background execution breaks the screenshot-analyze-act loop.

Run only one midscene command at a time. Wait for the previous command to finish, read the screenshot, then decide the next action. Never chain multiple commands together.

Allow enough time for each command to complete. Midscene commands involve AI inference and screen interaction, which can take longer than typical shell commands. A typical command needs about 1 minute; complex act commands may need even longer.

Always report task results before finishing. After completing the automation task, you MUST proactively summarize the results to the user — including key data found, actions completed, screenshots taken, and any relevant findings. Never silently end after the last automation step; the user expects a complete response in a single interaction.