How do a skill, repo instructions and a command that runs them differ?
Answer out loud or in your head first, then compare.
Answer
Repo instructions describe the project's general rules, e.g. how to run checks and where to find modules. A skill gathers the knowledge and steps needed for a particular kind of task. A command can be a way of running such a procedure.
These aren't always three separate technical mechanisms. In Claude Code, earlier custom commands and skills share a related model; check the current version of the tool. It's good when the context gets the materials needed now, not all the references at once. A skill isn't access protection. If it's meant to write data or run a command, the environment still has to control whether that's allowed.
Why it matters: Helps you talk about your own skill without mixing up the concepts.
Possible follow-ups: How do you check that a skill is triggered in the right situations?
How did it go? Be honest — “Not yet” puts the question on your review list.
Start with
- I-P01 · What is an agent harness? — Harness, MCP and coding agents
- I-P02 · What is MCP? — Harness, MCP and coding agents
- I-P04 · How does a coding agent differ from a chat that suggests code? — Harness, MCP and coding agents
Further reading
- Agent Skills: Specification(opens in a new tab)Spec; accessed 22.10.2026. SKILL.md, metadata, scripts, references and resources.
- Claude Code: Extend Claude with skills(opens in a new tab)Docs; accessed 22.10.2026. The current model of skills and custom commands; depends on the host.