This project addresses the challenge of synchronizing AI command-line interface (CLI) skills across multiple AI tools like Claude, Clawd Bot, and OpenCode. It provides a centralized "source of truth" for skills, allowing users to manage them efficiently from one location and sync them to all supported AI agents, benefiting developers and power users who leverage diverse AI CLI ecosystems.

How It Works

Skillshare establishes a single directory (~/.config/skillshare/skills/) as the master repository for all AI skills. A sync command then creates symbolic links from this master directory to the respective skill locations of various AI CLIs. This approach ensures that edits made in the master directory are immediately reflected across all integrated tools. The project also features an "AI-Native" component where skillshare itself acts as a skill that AI agents can invoke, enabling natural language commands for skill management and synchronization.

Quick Start & Requirements

• Primary install: curl -fsSL https://raw.githubusercontent.com/runkids/skillshare/main/install.sh | sh (manual) or brew install runkids/tap/skillshare (macOS).
• Prerequisites: A shell environment for installation. git is required for cross-machine synchronization and the Team Edition.
• Links: Installation script: https://raw.githubusercontent.com/runkids/skillshare/main/install.sh, Source code: https://github.com/runkids/skillshare.git, Documentation: docs/.

Highlighted Details

• "One source of truth for all your AI CLI skills. Edit once, sync everywhere."
• AI-Native integration allows AI agents to manage skills via natural language commands (e.g., "sync my skills").
• Supports bidirectional synchronization, enabling users to pull skill edits from any target agent back to the source.
• Team Edition offers features like tracked repositories for shared skills, nested skill structures, auto-pruning of orphaned links, and collision detection.

Maintenance & Community

The project is open for contributions via git clone and go build. Bugs and feature requests can be reported by opening an issue on GitHub. No specific community channels like Discord or Slack are mentioned in the README.

Licensing & Compatibility

The project is licensed under the MIT License, which is permissive and generally compatible with commercial use and closed-source linking.

Limitations & Caveats

Users are advised to run skillshare backup before adding custom targets that may contain existing files to prevent data loss. If skills do not appear after synchronization, running skillshare doctor and restarting the relevant CLI tool is recommended. Modifying skills within target directories directly edits the source due to the use of symlinks.