[Claude Code] How to Create and Use Skills - Your Own Slash Commands

If you're rewriting the same prompt every time, you can use the

This is what happens when you write Claude Code.

"Write a blog post. Oh, by the way, you should always write it in this style... Do I need to explain it again?"

It's annoying to have to type a long context like "write in Korean, in developer tone, with this structure" every time. And if you make a mistake, you end up with a completely different style.

This is where Skill comes in.

Skills are Claude Code's custom slash commands. You can save your favorite prompt patterns in a file, so you can recall them with a single short command like /blog-write. After creating and using it, I found myself thinking, "I wish I'd known about this a long time ago."

Describe the skill in one line

skill = markdown file with prompt template + slash command registration

No complicated setup, just create a file and it becomes a slash command.

Preparation

First, let's understand where skill files are stored.

~/.claude/skills/
└── skillname/
    └── skill.md ← This one file is the entirety of the skill.

~/.claude/skills/` is the global skills directory. It can be used in any project.
The skills folder name becomes the slash command name.
- ~/.claude/skills/blog-write/ → /blog-write
- ~/.claude/skills/gitpush/ → /gitpush

Step 1. Create the skill.md file

Let's create our own skill folder and files. For example, we'll create a simple commit message generator.

mkdir -p ~/.claude/skills/commit-msg

And create a file called ~/.claude/skills/commit-msg/skill.md.

---
name: commit-msg
description: Create a Korean conventional commit message based on git diff.
user-invocable: true
---

Analyze the current changes and write a conventional commit message in Korean.

## Rules
- Start with one of the following: `feat:`, `fix:`, `refactor:`, `docs:`, `style:`, or `chore:`.
- Title must be 50 characters or less
- If there are multiple files that have changed, summarize the key changes
- Keep the English technical terms, but explain in Korean

## Output format
Output only the commit message. No comments.

The file structure is divided into two main parts.

Delimiters	Content
Frontmatter (between `---`)	Skill meta information. Setting `name`, `description`, `user-invocable`
Body	The actual prompt. What will be delivered to Claude

Frontmatter field descriptions

---
name: commit-msg # Slash command name (/commit-msg)
description: one line description # description as seen in the list with /.
user-invocable: true # register to be able to write with the slash command
---

If user-invocable: true is not present, it will not appear in the slash list. You must include it.

Step 2. Use the skill

At the Claude Code prompt, type / to see a list of registered skills.

/commit-msg

After hitting enter, the body prompt of skill.md is automatically forwarded to Claude. It will analyze the current git diff and create a commit message right away.

Step 3. Get arguments with $ARGUMENTS

If you want to take additional input into your skill, you can use the $ARGUMENTS variable.

For example, if your blog post writing skill takes a topic as an argument:

---
name: blog-write
description: Write a developer blog post in Markdown.
user-invocable: true
---

Write a blog post on the following topics

## Style
- Korean, friendly and practical tone
- Include code examples
- Audience: First-time developers

## Topics
$ARGUMENTS

when used, type it like this

/blog-write How to create and use skills in Claude Code

"How to Create and Use Skills in Claude Code" is substituted for $ARGUMENTS.

Step 4. Make skill-creator a skill (easier way)

Claude Code has a built-in skill called skill-creator. This skill creates skills.

/skill-creator blog-write: Create a skill to write developer blog posts

Claude will listen to your skill description and automatically create the skill.md file for you. This is fast when you're not sure how to structure your prompts.

Summarize common patterns

1. Basic Skill Structure

---''
name: Skill name
description: One-line description
user-invocable: true
---.

Write the prompt here.

2. Skills that take arguments

---''
name: Skill name
description: Description (takes $ARGUMENTS arguments)
user-invocable: true
---.

topic: $ARGUMENTS

Please write ... with the above topic.

3. File path structure

~/.claude/skills/
├── blog-write/
│ └── skill.md
├── commit-msg/
│ └── skill.md
└── gitpush/
    └── skill.md

Skill vs. repetitive prompts

| | Type directly each time | Skill | |---|---|---|---| | Volume of input | Full long prompt | Single line of /skillname | | Consistency | Can vary every time | Always the same prompt | Arguments support | None | Can be done with $ARGUMENTS | | Sharing | Difficult | Can be shared as a file

Troubleshooting

Entering `/skillname` doesn't show up in the list

Make sure user-invocable: true is in frontmatter
Make sure the file path is ~/.claude/skills/skillname/skill.md (case sensitive)
Try restarting Claude Code

Prompt not working as intended

Unnecessary comments around $ARGUMENTS can confuse Claude
The more specific you are about the output format, the more stable the results will be
"Print ~" is more consistent than "Do ~ for me"

skill.md doesn't reflect after modification

After saving the file, it should be reflected when you start a new conversation in Claude Code.
It may not be reflected immediately in an existing conversation

Summary - Key flow at a glance

1. create a folder
   ~/.claude/skills/skillname/

2. create skill.md
   ---.
   name: Skill name
   description: Description
   user-invocable: true
   ---]
   [prompt content]
   [use $ARGUMENTS if necessary].

3. invoked with a slash in Claude Code
   /skillname [optional arguments].

If you have a repetitive task, it makes sense to turn it into a skill. It takes 5 minutes to create the first file, and after that it's just a single line of /skillname.

I wrote this blog post with the /blog-write skill. I just put in a topic and the structure and style are consistent. Once you get the hang of it, it gets easier and easier.