Extending Claude Code
Skills, hooks, subagents, MCP and plugins: what each one is for, and the difference between advice and a rule.
6 chapters · 23 min
Start the course- 1Skills, loaded only when neededA skill is a folder whose description sits in context and whose body arrives only when the task calls for it. That split is what makes skills cheap, and what makes a bad description expensive.4 min
- 2Hooks turn a wish into a guaranteeA CLAUDE.md line is a request the model usually honours. A hook is code the harness runs, so it happens every time. Choose hooks only for rules that have no exceptions.4 min
- 3Subagents keep the mess out of your sessionA subagent works in its own context window and returns only a summary. That is the whole benefit, and it is also the whole limit, so it pays to know what comes back.4 min
- 4MCP connects tools, and every tool costs contextMCP gives Claude tools that live outside your machine's built-ins. The setup is one command. The real decisions are which scope to use and whether a plain CLI would have done the job for free.4 min
- 5Plugins package what already worksA plugin is not a new capability. It bundles skills, subagents, hooks and MCP servers into one installable, versioned unit, and an enabled plugin costs context every turn whether you use it or not.3 min
- 6Advice or rule: choosing the right toolCLAUDE.md, skills, hooks, subagents and MCP each answer a different question. Pick by asking what must be true, how often it is needed, and what you are willing to carry in context.4 min