Create Command Skill

Create new Claude Code slash commands with proper configuration and advanced features.

Command Types

Project Commands

Commands stored in your repository and shared with your team.

Location: .claude/commands/ or plugins/<plugin>/commands/

Shows "(project)" in /help.

Personal Commands

Commands available across all your projects.

Location: ~/.claude/commands/

Shows "(user)" in /help.

Command File Format

Commands are markdown files with YAML frontmatter:

---
description: Short description shown in /help
argument-hint: <required-arg> [optional-arg]
---

# Command Name

Instructions for Claude when this command is invoked...

Frontmatter Fields

Field	Purpose	Default
description	Brief description of the command	Uses first line from prompt
argument-hint	Arguments shown in autocomplete	None
allowed-tools	List of tools the command can use	Inherits from conversation
model	Specific model string	Inherits from conversation
disable-model-invocation	Prevent SlashCommand tool from calling	false

Argument Placeholders

All Arguments: $ARGUMENTS

Captures all arguments passed to the command:

---
description: Fix a GitHub issue
argument-hint: <issue-number> [priority]
---

Fix issue #$ARGUMENTS following our coding standards.

# Usage: /fix-issue 123 high-priority
# $ARGUMENTS becomes: "123 high-priority"

Positional Arguments: $1, $2, etc.

Access specific arguments individually:

---
description: Review pull request
argument-hint: <pr-number> <priority> <assignee>
---

Review PR #$1 with priority $2 and assign to $3.
Focus on security, performance, and code style.

# Usage: /review-pr 456 high alice
# $1 = "456", $2 = "high", $3 = "alice"

Bash Command Execution

Execute bash commands before the slash command runs using the ! prefix.

IMPORTANT: You MUST include allowed-tools with the Bash tool.

---
allowed-tools: Bash(git add:*), Bash(git status:*), Bash(git commit:*)
description: Create a git commit
---

## Context

- Current git status: !`git status`
- Current git diff: !`git diff HEAD`
- Current branch: !`git branch --show-current`
- Recent commits: !`git log --oneline -10`

## Your task

Based on the above changes, create a single git commit.

File References

Include file contents using the @ prefix:

---
description: Review implementation
---

# Reference a specific file
Review the implementation in @src/utils/helpers.js

# Reference multiple files
Compare @src/old-version.js with @src/new-version.js

Namespacing

Organize commands in subdirectories. The subdirectory appears in the description but not the command name.

.claude/commands/
├── frontend/
│   └── component.md    # Creates /component (project:frontend)
└── backend/
    └── api.md          # Creates /api (project:backend)

Argument Hint Conventions

Syntax	Meaning
<name>	Required argument
[name]	Optional argument
[--flag]	Optional flag
[--option <value>]	Optional flag with value
"" or omit	No arguments

Examples

Simple Command

---
description: Show project status overview
---

Show a summary of the current project status including:
- Git branch and recent commits
- Open issues and PRs
- Build status

Command with Tools

---
allowed-tools: Bash(npm:*), Bash(yarn:*)
description: Run tests with coverage
argument-hint: [--watch] [--coverage]
---

Run the test suite with the following options:
- If --watch is passed, run in watch mode
- If --coverage is passed, generate coverage report

Use npm or yarn based on the lock file present.

Git Commit Command

---
allowed-tools: Bash(git add:*), Bash(git status:*), Bash(git commit:*)
description: Create a conventional commit
argument-hint: [message]
---

## Context

- Current status: !`git status --short`
- Staged changes: !`git diff --cached --stat`
- Recent commits: !`git log --oneline -5`

## Task

Create a git commit following conventional commit format:
- type(scope): description
- Types: feat, fix, docs, style, refactor, test, chore

If $ARGUMENTS is provided, use it as the commit message.
Otherwise, analyze the changes and generate an appropriate message.

Deploy Command

---
allowed-tools: Bash(docker:*), Bash(kubectl:*)
description: Deploy to environment
argument-hint: <environment> [--dry-run]
model: claude-sonnet-4-20250514
---

Deploy to $1 environment.

## Pre-flight checks
- Verify current branch: !`git branch --show-current`
- Check for uncommitted changes: !`git status --porcelain`

## Process
1. Build Docker image
2. Push to registry
3. Update Kubernetes deployment
4. Verify rollout status

If --dry-run is in $ARGUMENTS, only show what would be done.

Code Review Command

---
description: Review code changes
argument-hint: [file-path]
---

Review the code for:
- Security vulnerabilities
- Performance issues
- Code style violations
- Best practices

If a file path is provided ($1), focus on that file.
Otherwise, review all changed files:
!`git diff --name-only HEAD~1`

Reference style guide: @.claude/STYLE_GUIDE.md

Thinking Mode

Include extended thinking keywords to trigger deeper analysis:

---
description: Analyze architecture deeply
---

Think step by step about the architecture of this codebase.
Consider:
- Design patterns used
- Potential improvements
- Scalability concerns

SlashCommand Tool Integration

For commands to be invocable by Claude via the SlashCommand tool:

1. Must have description frontmatter field
2. Use disable-model-invocation: true to prevent automatic invocation
3. Built-in commands (like /compact) are NOT supported

Permission Rules

SlashCommand:/commit           # Exact match only
SlashCommand:/review-pr:*      # Prefix match with any arguments

File Location

Type	Location
Plugin commands	plugins/<plugin>/commands/<name>.md
Project commands	.claude/commands/<name>.md
Personal commands	~/.claude/commands/<name>.md

The filename (without .md) becomes the slash command name.

Slash Commands vs Skills

Aspect	Slash Commands	Skills
Complexity	Simple prompts	Complex capabilities
Structure	Single .md file	Directory with SKILL.md + resources
Discovery	Explicit (/command)	Automatic (context-based)
Files	One file only	Multiple files, scripts, templates

Use slash commands when:

• You invoke the same prompt repeatedly
• The prompt fits in a single file
• You want explicit control over when it runs

Use Skills when:

• Claude should discover automatically
• Multiple files or scripts needed
• Complex workflows with validation