Module 5⬡ The Crossover

The Commander

You leave this module with a permanent AI harness: a CLAUDE.md for your domain, slash commands for your recurring workflows, and the proven ability to run autonomous AI sessions. The harness improves every time you add a convention. It compounds forever.

Estimated time: 6–8 hrsPrimary tool: Claude Code (Haiku/Sonnet)9 drills

Agentic engineering. You are not writing code. You are not even describing code. You are defining outcomes, providing context, setting constraints, and delegating execution. The AI is the pilot. You are mission control.

⬡The Crossover

This module contains The Crossover — three drills that permanently change how you relate to AI. Drills 5.3, 5.4, and 5.7 are the tipping point sequence. Everything before this module is learning to use AI. This module is where AI becomes yours. Module 6 makes it immortal.

Stage 1 of 3: The Seed

AI stops feeling like a stranger and starts feeling like a colleague who was already in the room.

Stage 2 of 3: The Muscle

A recurring workflow becomes a single word. You're no longer using AI — you're configuring it.

Stage 3 of 3: The Proof

Walk away. Come back. Work is done. This is not a tool you operate. This is capacity you deploy.

Drills in this module

5.1

The Install

After this drill, Claude Code is installed, your API key is configured, and you have run your first command.

20 min

5.2

The First Task

After this drill, you have delegated a well-scoped task to Claude Code and seen it execute — reading code, making changes, and reporting back.

25 min

5.3⬡

The CLAUDE.md

After this drill, you have written project memory for your app — and for the first time, Claude Code already knows your context without being told.

30 min

5.4⬡

The Slash Command

After this drill, you have created a slash command that compresses a recurring workflow into a single word — and felt the muscle memory of a personalized AI.

30 min

5.5

The Model Ladder

After this drill, you have run the same task across Haiku, Sonnet, and Opus and can make cost-vs-quality trade-offs consciously.

25 min

5.6

The Plan Mode

After this drill, you use plan mode as a habit — reviewing every complex execution before authorizing it.

25 min

5.7⬡

The Autonomous Session

After this drill, you have walked away from the keyboard while Claude Code worked, returned to find the work done, and understood in your bones: this is not a tool I operate. This is capacity I deploy.

30 min (15 away from keyboard)

5.8

The Mission

After this drill, you have run a complete MCPP-lite cycle — Intent → Plan → Execute → Report — and understand the pattern behind all effective agentic engineering.

40 min

5.9

The Harness Build

After this drill, you have built your own AI harness — a CLAUDE.md and 3 slash commands for your specific domain — ready to deploy on any future project.

45 min

Module sample project

The Agent-Built Feature

Use Claude Code to add a significant feature to your project from Module 4. Write a mission prompt using MCPP-lite, review the plan, authorize execution, and verify results. The AI does the implementation. You do the commanding.

·A new feature implemented by Claude Code (not written by you)
·Your CLAUDE.md file (the project memory from Drill 5.3)
·At least 2 working slash commands (.claude/commands/)
·MCPP-lite mission brief for the feature
·A brief retrospective: what Claude Code did, what you corrected

Key lesson

“The quality of AI output is not a function of the model's intelligence. It is a function of the context you engineer and the authority you delegate. This is agentic engineering.”