A suite of automation skills that allow AI agents to control macOS via Accessibility APIs, handling everything from window management to OCR-based interactions.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install macpilot
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install macpilot using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is MacPilot Skills?
MacPilot Skills provide a robust framework for AI coding agents to interact directly with the macOS operating system. By utilizing the MacPilot CLI and macOS Accessibility APIs, these Openclaw Skills enable agents to perform actions that are typically restricted to human users, such as navigating complex GUI menus, interacting with native file dialogs, and managing system-level configurations. This integration bridges the gap between terminal-based coding and full desktop automation.
Whether you are using Claude Code, Cursor, or other compatible tools, these skills empower your agent to see and interact with your Mac environment. The focus is on reliability and precision, ensuring that the agent can find, click, and modify UI controls across any application without manual intervention.
MacPilot Skills Use Cases
- Automating repetitive UI-based workflows across different macOS applications.
- Handling native file operations like saving, opening, or printing documents through system dialogs.
- Orchestrating developer workspaces by automatically snapping and resizing windows into specific layouts.
- Performing screen-based data extraction using OCR to click on text-based buttons or read labels.
- Managing system state through clipboard history, notifications, and shell execution.
How MacPilot Skills Works
- The AI agent loads the Openclaw Skills into its environment, gaining access to specialized tools for macOS interaction.
- When a task is requested, the agent invokes MacPilot commands to query the macOS Accessibility API for the current UI state.
- MacPilot identifies target elements, windows, or system controls based on the agent's intent.
- The skill executes the specified action, such as a mouse click, keyboard input, or window snap, and returns the visual or state feedback to the agent.
- The agent verifies the outcome through UI inspection or screenshots to ensure the automation task was completed successfully.
MacPilot Skills Setup
First, ensure MacPilot is installed and that you have granted Accessibility and Screen Recording permissions in System Settings > Privacy & Security. To install the Openclaw Skills, use the following command:
npx skills add adhikjoshi/macpilot-skills
For manual installation into a specific agent like Claude Code, create a skills directory and copy the relevant markdown files:
mkdir -p .claude/skills/macpilot-automation
cp skills/macpilot-automation/SKILL.md .claude/skills/macpilot-automation/
MacPilot Skills Data Schema & Taxonomy
The skill organizes its interactions by categorizing system elements and metadata into structured formats for the agent. This includes:
| Category | Data Handled |
|---|---|
| UI Elements | Roles, descriptions, and coordinate mapping for accessible controls. |
| Window Management | Position, size, snap state, and workspace layout configurations. |
| Screen Capture | OCR text results, screen recording segments, and visual verification frames. |
| System Context | Clipboard history, notification contents, and active application lists. |
MacPilot Skills Advanced Features
- Advanced OCR-click capabilities to interact with non-standard UI elements by finding text on screen.
- Deep integration with native macOS dialogs (Open, Save, Print) including folder navigation.
- Comprehensive window management with support for Spaces and layout restoration.
- Clipboard history search and management for complex data handling between apps.
- Multi-application orchestration through centralized accessibility controls.
SKILL.md
Loading
Related Openclaw Skills
Halocard Virtual Cards
Generate single-use virtual credit cards on the fly to complete secure online payments during automated agent tasks.
MenuVision
MenuVision transforms restaurant menus from any source into interactive, visually rich HTML digital menus using advanced AI vision and image generation.
Grok Imagine Extended
A comprehensive tool for generating high-fidelity images and cinematic videos using xAI's Grok Imagine Extended models.
openclaw-daily-tips
A daily optimization engine that delivers actionable strategies to improve AI agent performance and reduce operational costs.
Jina AI Reader & Search
A comprehensive toolset for converting web pages to LLM-friendly markdown and performing multi-step deep research.
Maps
A location-intelligence skill providing geocoding and routing capabilities without requiring external API keys.
Featured*
