mogkit

Setup 101

If you've never opened a terminal,
start here.

mogkit runs in your computer's terminal. The terminal is just a text-based way to talk to your computer — it looks intimidating because of the movies, but the moves are simple. We'll go step by step. By the end of this page you'll have run your first PM skill.

Estimated time: 10–15 minutes, most of it waiting for installs.

Step 1

Open the terminal

Open the app called Terminal:

  • Mac: press ⌘ + Space, type Terminal, hit Enter.
  • Windows: install WSL first (one-time), then open the Ubuntu app.
  • Linux: you already know.

You'll see a blinking cursor next to your username. That's the prompt. You'll type a command, hit Enter, and the computer responds. That's the whole loop.

Step 2

Install Node 20+

Node is the JavaScript runtime mogkit's CLI runs on. Check whether you already have it: 

node --version

If that printed v20.x.x or higher, you're done with this step. If it said "command not found" or showed a version below 20, install Node:

  • Easy: download the LTS installer from nodejs.org.
  • Better: install nvm and run nvm install 20. nvm lets you manage Node versions without a fight.

Step 3

Install Claude Code

Claude Code is the agent that runs the mogkit skills. Install it per the official instructions at claude.com/claude-code — it ships as a CLI you'll launch from inside your workspace.

Once installed, you should be able to run:

claude --version

and see a version string come back.

Step 4

Install mogkit + scaffold a workspace

Install the CLI globally — this puts mogkit on your PATH so every subsequent command works: 

npm install -g mogkit

Then pick a folder you keep projects in and scaffold: 

mogkit init my-workspace
cd my-workspace

That's it. You now have a folder with:

  • CLAUDE.md — orients every Claude Code session in this workspace
  • sources/ — your research corpus, version-controlled
  • graph/ — generated by the Discovery skills
  • knowledge/ — outputs of interrogation skills land here
  • .claude/skills/ — all 13 mogkit skills installed and ready

Step 5

Check your setup

From inside the workspace, run:

mogkit doctor

Every check should pass with a green ✓. If something fails, the message tells you what to fix and how. No googling required.

Want to play with a real corpus right away? Drop in our sample: 

mogkit demo

That loads 8 mixed-type sources (interviews, tickets, a PRD, a memo, competitor research) into sources/ — enough to see every Discovery skill actually do something on day one.

Step 6

Run your first skill

Open the workspace in Claude Code and try a standalone skill. A good first one: tradeoff-frame. Paste in a contested decision you're chewing on at work right now — two real options. The skill will return the axes of the disagreement, what each option optimizes vs sacrifices, and whether the call is a one-way or two-way door.

It won't pick the answer. That's deliberate. Frames the decision so you can decide well.

Where to go next

  • The Discovery wedge — mogkit's deepest system. Worth setting up first if your work is research-led.
  • The Workflow Library — implementable AI-for-PM workflows. Each one teaches its mechanism.
  • If anything sticks, run mogkit doctor — it diagnoses most setup issues in plain language.