You'll finish with: Your first working /command.
Skills are a Claude Code feature, and you said you don't use it — so you can skip this one.
A skill is a saved procedure you fire with a slash command. It needs Claude Code in a terminal. The nearest thing you already have is your project's Instructions, which apply to everything you do in that project rather than waiting to be called by name.
If you find yourself typing the same instructions repeatedly, add them to your Instructions box instead. That gets you most of the benefit with none of the setup.
You can name one thing you type over and over. Add it to your project Instructions and it stops being something you type.
If you can't think of one, you don't need this — genuinely. Don't build a skill for a job you do twice a year.
You've typed the same instructions into chat three times this month. That's a skill.
A skill is a file containing a procedure. Claude loads it when relevant, or you fire it by typing /name.
What a skill costs you
Only the one-line description sits in context permanently. The body loads when used. A long playbook costs almost nothing until you need it.
~/.claude/skills/quote/SKILL.md inside it. The folder name becomes the command./quote.---
description: Builds a client quote from a job description.
Use when asked to quote a job or price up work.
disable-model-invocation: true
---
Build a quote for: $ARGUMENTS
1. Map the job to an offer in business/offers.md
2. If it doesn't map cleanly, stop and ask
3. List what's in scope, and what's explicitly out
4. Give the range from offers.md, not a fixed number
5. List every assumption you made
Draft only. Never send. Never invent a price.Type /quote and it runs your procedure — not a generic answer.
If the command doesn't exist, the folder name is wrong. The folder becomes the command, so skills/quote/SKILL.md gives you /quote.
disable-model-invocation: true means only you can run it — Claude won't decide on its own that now's a good time to build a quote. Use it on anything with a consequence.