M06 · Claude Code (CLI)

Stage C — Tools · Persona: Operators + Builders · Duration: 120 min · Format: live / hands-on · Prereqs: M05

Objective

Get participants comfortable using Claude Code in the terminal to do real, multi-step work — editing files, running commands, and completing an actual company task — not just chatting.

Why it matters

Claude Code is where AI stops being a chat box and becomes an agent that does things on your machine and your files. It's the gateway to everything in Stages D–F, and it's how the OS gets built and operated.

Learning outcomes

By the end, participants can: - Install and launch Claude Code and understand what it can/can't touch. - Give it a multi-step task and supervise it (read its plan, approve actions). - Work with files and run commands through it safely (permissions, what to allow). - Use it for non-coding work too (documents, data, research, ops). - Complete one real company task end-to-end in the terminal.

Agenda

1. From chat to agent: what Claude Code is (10)
2. Install, launch, first task (20)
3. How it works: reading files, editing, running commands, permissions (25)
4. Beyond code: docs, data, ops use cases (15)
5. Lab: each participant completes one real task (40)
6. Safety, habits, and where this leads (10)

Deliverable

One real company task completed in Claude Code → deliverable.md, with a short write-up of what was done.

Gate contribution

Gate 3 (Building) — alongside M07/M08.