Runs Claude Skills locally on macOS with Apple Silicon, enabling private, offline processing of documents, spreadsheets, and images. This tool provides a sandboxed environment via an MCP server, allowing users to leverage specialized AI skills without uploading sensitive data, offering enhanced privacy and control. It targets developers and power users seeking to integrate local AI capabilities with their existing workflows.

How It Works

OpenSkills deploys a local MCP (Model Communication Protocol) server that acts as a compatibility layer. It utilizes Apple's native container technology to create a VM-level isolated sandbox environment. AI-generated code and scripts from Claude Skills are executed within this secure sandbox, enabling local file processing. This architecture ensures that data remains on the user's Mac, providing a secure and private execution context for AI-driven tasks.

Quick Start & Requirements

• Primary Install: Clone the repository, navigate to the directory, make install.sh executable, and run it (chmod +x install.sh && ./install.sh). Installation takes approximately 2 minutes.
• Prerequisites: macOS with Apple Silicon (M1/M2/M3/M4/M5), Python 3.10+. A requirements.txt file is available for additional Python dependencies.
• MCP Server: The local MCP server becomes available at http://open-skills.local:8222/mcp.
• Integration: Configure AI tools like Claude Desktop, Gemini CLI, Qwen CLI, or Python OpenAI Agents by updating their settings to point to the local MCP server URL.
• Links:
  • GitHub Repository: github.com/BandarLabs/open-skills
  • Anthropic Skills: github.com/anthropics/skills
  • Skills Documentation: docs.claude.com/skills

Highlighted Details

• Privacy-First: Processes all user data locally on the Mac, eliminating the need for cloud uploads.
• VM-Level Isolation: Leverages @apple/container for robust security, ensuring the host system remains protected.
• Broad Compatibility: Works with any MCP-compatible tool, including Claude Desktop, Gemini CLI, Qwen CLI, Kiro, and Python OpenAI Agents.
• Extensible Skill System: Supports importing Anthropic's official skills or creating custom skills locally.

Maintenance & Community

The project encourages contributions and issue reporting via its GitHub repository. Specific community channels like Discord or Slack, or details on notable contributors and sponsorships, are not detailed in the README.

Licensing & Compatibility

• License: Apache 2.0.
• Compatibility: The permissive Apache 2.0 license generally allows for commercial use and integration into closed-source projects, though specific AI tool integrations may have their own terms.

Limitations & Caveats

This tool is strictly limited to macOS environments equipped with Apple Silicon processors (M1/M2/M3/M4/M5); it does not support Intel-based Macs or other operating systems.