An AI slash command that writes the tutorial, not the code.
AI coding agents are powerful — but they make it easy to stop thinking. You ask for a feature, the agent writes it, you paste it in. You ship it. You never really understood what happened.
/lmf flips that dynamic. Instead of writing the code for you, the agent
writes a step-by-step how-to guide — and then gets out of the way.
You read it, you understand it, you write it yourself.
Prefix any task with /lmf and the agent will produce a Markdown tutorial
instead of code. The file is saved locally so you can follow it at your
own pace, offline, without the agent in the loop.
/lmf set up JWT authentication in a Laravel API
Produces: howto/0001_set-up-jwt-authentication-in-a-laravel-api.md
The tutorial covers what to do, why each step matters, what to watch out for — all in plain prose. No code blocks. No copy-paste shortcuts.
npx skills add radenadri/lmf
Every tutorial generated by /lmf follows this structure:
| Section | What It Contains |
|---|---|
| Overview | What you will build and why it matters |
| Prerequisites | What you need before starting |
| Steps | Prose instructions with reasoning for each action |
| Checkpoint | How to verify everything is working |
| Common Pitfalls | What usually goes wrong and how to fix it |
Files are saved to howto/ relative to your current working directory,
with auto-incrementing 4-digit numbering:
howto/
├── 0001_set-up-jwt-authentication-in-a-laravel-api.md
├── 0002_configure-redis-as-a-queue-driver-in-docker.md
└── 0003_add-role-based-access-control-to-next-js.md
If a file already exists, the agent will ask for confirmation before overwriting it.
- No code blocks — not even one line
- No inline code where plain prose works
- No copy-paste shortcuts — describes what to type and why
- Every step explains the reasoning, not just the action
- Written for someone who knows programming but is new to this specific topic
The goal is not to avoid AI. It is to use AI as a teacher, not a ghostwriter.
/lmf is for the times you want to actually understand what you are
building — so that next time, you do not need to ask at all.
MIT