7 Ways to Control Claude Code (Most People Only Know One)

So, an introductive cheat sheet for controlling Claude Code. You can customize it. CLAUDE.md and files, skills, hooks, rules, subagents, output styles, and appending the system prompt. When an animation looks into context. When a table of contents is used for a project overview. And you much authority. In this video, I'll break down the summary of key differences across each method while the post-prompt additional default is the most important for determining what breath of your Claude instructions to manage. Here's all seven. Okay, one CLAUDE.md. This is the always on table of contents view of your project. I covered this in my last video. Number two, rules. This is specific do's and don'ts that can load when Claude touches certain files. Number three, skills. These are step-by-step playbooks that only load when you trigger them. And number four, subagents. You can think of these as like little helpers. They go and do a side task in a new window and then report back with the answer. Number five, hooks. These are automations that fire on their own, like pinging slack when it's done. Number six, output styles. These change Claude's entire role, like switching from a coder to a general assistant. And last one, appending the system prompt. So these are quick add-on instructions for tone or formatting only for that one session. So every one of these controls, when an instruction shows up, and how much weight it carries. And that's how you stop overloading Claude and actually start directing it. If you comment control, I'll send you my full breakdown on all seven.