Model choice is the single largest lever you have over what Claude Code costs. It is also the setting most people never touch, because the default works and nobody told them there was a decision to make. This module makes it a decision. You will learn the alias table, the four-step ladder that determines which model actually runs, how effort levels change the bill, and which token you are really paying for.

A skill is how you teach Claude Code a procedure once and stop explaining it. It is also the cheapest surface you have, because a skill's body does not cost you anything until the skill is used. This module covers what a skill is, what its frontmatter can override, and how to fork a skill to a cheaper model so the work you repeat most often is the work you pay least for.

Two ways to extend Claude Code beyond one conversation. Subagents give you more workers, each in its own context with its own limits. MCP gives you more reach, connecting Claude Code to the systems your business already runs on. This module covers both, and both are as much about constraint as about capability.

Everything up to this point has been instructions and configuration. This module is about enforcement. A hook is the only surface that fires whether Claude agrees with it or not, which makes it the only surface that can actually guarantee anything. You will learn the lifecycle, the exit-code contract, how to write a guard that blocks a dangerous command outright, and how to think about hooks as governance rather than as scripting.

You now have five surfaces worth of mechanics. This module turns them into one posture. You will see all of it assembled for a single small business, work out what Delegate-level responsibility actually asks of you, and learn to recognize the point where a job has outgrown Claude Code.