I've been feeling the itch to divide my skills repo into two kinds of skills: - Model-invocable skills (skills) - User-invocable skills (commands) I.e. you'd be able to run /improve-codebase-architecture (a command), which would use /deep-modules (a skill) You'd run /grill-with-docs, which relies on /domain-modeling That way, skills become more composable. You can ask "give me a report on the /deep-modules in this repo", or "help me with /domain-modelling a new concept"