add-skill

Adding new skills involves creating skill directories, defining frontmatter, and setting up symlinks for your specific AI platform. This process ensures the assistant correctly recognizes and integrates the new capabilities for either global or project-specific use.

Global skills are managed through dotfiles and apply universally across your environment, while project-local skills are specific to a single repository. The integration process differs based on which scope you choose for your AI assistant.

Setting up symlinks for AI platforms like Claude and Codex involves creating specific directory links during the integration process. This connects your skill files and frontmatter to the platform's recognized paths so they can be accessed.

You can verify that a new skill is correctly recognized by your AI system by running stowing and path checks. These commands confirm that the newly created directories and symlinks are properly configured and accessible.

Yes, you can create repo-specific skills that are not managed by dotfiles. The workflow outlines how to set up project-local capabilities that remain entirely independent of your global AI assistant configuration.

Yes, defining frontmatter is a required step when creating a new skill directory. It provides the necessary metadata for the AI assistant to properly read, process, and integrate the skill into its system.