GitHub CLI Launches "gh skill" Command to Manage AI Agent Skills

GitHub CLI introduces "gh skill" today, a new command reshaping how developers work with AI coding agents. This tool makes it easy to discover, install, manage, and publish agent skills directly from GitHub repositories. The developer community now has a standardized and efficient method to empower their intelligent assistants.

What exactly are agent skills? They are portable sets of instructions, scripts, and resources that teach AI agents how to perform specific tasks. These skills follow the open "Agent Skills" specification. They also work across multiple agent hosts, including GitHub Copilot, Claude Code, Cursor, Codex, and Gemini CLI, among others.

Getting started with "gh skill" is straightforward. First, ensure you update your GitHub CLI to version v2.90.0 or later. Once updated, you can start discovering and installing skills interactively right from your command line.

• To browse skills in a repository and install them interactively, use: gh skill install github/awesome-copilot
• If you want to install a specific skill directly, try: gh skill install github/awesome-copilot documentation-writer
• To install a specific version, indicate the tag: gh skill install github/awesome-copilot [email protected]
• You can also install a skill at a specific commit SHA: gh skill install github/awesome-copilot documentation-writer@abc123def
• And to discover skills, simply search with: gh skill search mcp-apps

Skills are automatically installed to the correct directory for your agent host. You also have the flexibility to target a specific agent and scope using flags. For instance, gh skill install github/awesome-copilot documentation-writer --agent claude-code --scope user directs the installation.

Supply chain integrity is a critical aspect. Agent skills are executable instructions that shape how AI agents behave. A skill that changes silently between installs poses a significant supply chain risk. "gh skill" brings the same guarantees you expect from package managers to the skills ecosystem. It leverages primitives GitHub already provides to ensure trust and transparency.

• Tags and releases: Every published release is tied to a Git tag. "gh skill publish" offers to enable immutable releases. This means release content cannot be altered after publication, even by administrators.
• Content-addressed change detection: Each installed skill records the Git tree SHA of its source directory. "gh skill update" compares local SHAs against the remote. This detects real content changes, not just version bumps. By storing this information in skills' "frontmatter", versioning and pinning are portable too. This allows you (or your agent) to copy and paste the skill to different projects without losing the ability to track changes and update it.
• Version pinning: You can lock a skill to a specific tag or commit SHA with the "--pin" flag. Pinned skills are skipped during automatic updates. This ensures deliberate upgrades, not accidental ones.
• Portable provenance via frontmatter: When "gh skill" installs a skill, it writes tracking metadata. This includes the repository, reference, and tree SHA. This information is stored directly in the "SKILL.md" "frontmatter". Provenance data travels with the skill, no matter where it ends up. Users, agents, and scripts frequently move, copy, and reorganize skills.