Creating Slash Commands

This Skill provides comprehensive knowledge about creating effective slash commands in Claude Code.

What Are Slash Commands?

Slash commands are user-invoked operations that start with /. When executed, they expand to a full prompt that Claude processes. They're different from subagents (which are automatically invoked) - slash commands require explicit user action.

File Structure

Project commands: .claude/commands/{command-name}.md Personal commands: ~/.claude/commands/{command-name}.md

Format:

---
description: Brief description of what this command does
argument-hint: [arg1] [arg2]  # Optional
allowed-tools: Bash(git:*), Read, Write  # Optional
model: haiku  # Optional
disable-model-invocation: false  # Optional
---

Command prompt that Claude will execute.
Can use $ARGUMENTS or $1, $2, $3, etc.

YAML Frontmatter Fields

Optional Fields (All Are Optional)

description (string)

• Brief description shown in /help
• If omitted, uses first line of prompt
• Keep concise (one sentence)

argument-hint (string)

• Shows expected arguments in autocomplete
• Examples: [filename], [message], add [tag] | remove [tag]
• Helps users know what to provide

allowed-tools (comma-separated string)

• Tools this command can use
• If omitted, inherits from conversation
• Use for command-specific restrictions

model (string)

• Specific model: sonnet, opus, haiku
• If omitted, inherits from conversation
• Use haiku for simple/fast commands

disable-model-invocation (boolean)

• If true, prevents SlashCommand tool from calling this command
• Default: false
• Use to keep commands user-only

Command Arguments

Using $ARGUMENTS (All Arguments)

Captures everything after the command name:

---
description: Fix issue following coding standards
---

Fix issue #$ARGUMENTS following our coding standards

Usage:

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

Using $1, $2, $3 (Positional Arguments)

Access specific arguments individually:

---
description: Review PR with priority and assignee
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"

When to Use Which

Use $ARGUMENTS:

• Flexible, variable-length input
• Natural language descriptions
• Unknown number of args

Use $1, $2, $3:

• Fixed structure
• Specific fields needed
• Build complex prompts with arg positioning

Bash Command Execution

Execute bash commands before the prompt runs using ! prefix.

Basic Execution

---
allowed-tools: Bash(git:*)
---

## Current Status

Git status: !`git status`

## Your Task

Based on the above status, create a commit.

The ! commands run first, output is included in prompt context.

Multiple Commands

---
allowed-tools: Bash(git:*), Bash(docker:*)
---

## Context

- Git status: !`git status`
- Docker status: !`docker ps`
- Current branch: !`git branch --show-current`

## Task

Deploy the current branch.

Required: allowed-tools

Must include Bash tool permissions:

---
allowed-tools: Bash(git status:*), Bash(git diff:*)
---

Current changes: !`git diff HEAD`

Specify exact commands or use wildcards:

• Bash(git:*) - All git commands
• Bash(git status:*), Bash(git diff:*) - Specific commands only

File References

Include file contents using @ prefix:

---
description: Review implementation
---

Review the implementation in @src/utils/helpers.js

Compare @src/old-version.js with @src/new-version.js

When command runs, files are automatically included in context.

Extended Thinking

Trigger extended thinking by including keywords:

---
description: Solve complex algorithm problem
---

Think carefully about this algorithm optimization problem: $ARGUMENTS

Consider multiple approaches before implementing.

Keywords that trigger thinking: "think carefully", "consider", "analyze deeply"

Command Patterns

Simple Action Commands

---
description: Run all tests
allowed-tools: Bash(pytest:*)
---

Run the complete test suite:
!`pytest -v`

Report results and any failures.

Context Gathering Commands

---
description: Analyze current work
allowed-tools: Bash(git:*)
---

## Current Work Context

- Branch: !`git branch --show-current`
- Status: !`git status --short`
- Recent commits: !`git log --oneline -5`
- Uncommitted changes: !`git diff HEAD --stat`

Summarize what I'm currently working on.

Workflow Commands

---
description: Create feature branch from issue
argument-hint: [issue-number]
allowed-tools: Bash(gh:*), Bash(git:*)
---

## Issue Details

!`gh issue view $1 --json title,body`

## Task

1. Create branch: `issue-$1-{description}`
2. Check out the branch
3. Summarize the issue for me

Validation Commands

---
description: Validate code before commit
allowed-tools: Bash(git:*), Bash(npm:*), Bash(pytest:*)
---

## Pre-Commit Checks

1. Run linter: !`npm run lint`
2. Run tests: !`pytest`
3. Check types: !`npm run type-check`
4. Git status: !`git status`

Report any issues that need fixing before commit.

Generation Commands

---
description: Generate API endpoint
argument-hint: [resource-name]
---

Create a complete REST API endpoint for: $1

Include:
- Model definition
- Schema (Create, Update, Response)
- CRUD operations
- Endpoints (GET, POST, PUT, DELETE)
- Tests

Tool Restrictions

Git Commands Only

---
allowed-tools: Bash(git:*), Read, Write
---

Safe git operations command.

Read-Only Command

---
allowed-tools: Read, Grep, Glob
---

Analysis command that doesn't modify files.

Specific Commands

---
allowed-tools: Bash(docker compose up:*), Bash(docker compose down:*)
---

Docker management command with limited operations.

Model Selection

Haiku for Simple Commands

---
model: haiku
---

Quick file listing: !`ls -la`

Fast, economical for simple tasks.

Sonnet for Standard Commands

# Omit model field - uses conversation model (usually sonnet)
---
description: Standard complexity command
---

Opus for Complex Commands

---
model: opus
---

Analyze architecture and suggest improvements: $ARGUMENTS

Deep reasoning for complex tasks.

Examples

Example 1: Simple Git Commit

---
description: Create git commit with message
argument-hint: [message]
allowed-tools: Bash(git:*)
---

Create a git commit:

Message: $ARGUMENTS

Steps:
1. Stage all changes: `git add .`
2. Commit with message
3. Show commit hash and summary

Example 2: Test Runner

---
description: Run specific test file
argument-hint: [test-file]
allowed-tools: Bash(pytest:*)
---

Run tests in: $1

!`pytest $1 -v`

Analyze failures and suggest fixes.

Example 3: Context-Heavy Command

---
description: Prepare for code review
allowed-tools: Bash(git:*), Read
---

## Code Review Preparation

### Changes
!`git diff main..HEAD --stat`

### Commits
!`git log main..HEAD --oneline`

### Modified Files
!`git diff main..HEAD --name-only`

Summarize changes for code review and suggest review focus areas.

Example 4: Branching Workflow

---
description: Create feature branch from issue
argument-hint: [issue-number]
allowed-tools: Bash(gh:*), Bash(git:*)
---

## Issue #$1 Details

!`gh issue view $1 --json title,body,labels`

## Task

1. Create branch: `issue-$1-{short-description}`
2. Switch to branch
3. Print next steps for working on this issue

Example 5: Multi-Step Workflow

---
description: Complete issue and create PR
argument-hint: [issue-number]
allowed-tools: Bash(git:*), Bash(gh:*)
---

## Current Status

Branch: !`git branch --show-current`
Status: !`git status --short`

## Task

Complete issue #$1 workflow:
1. Ensure all changes committed
2. Push branch to origin
3. Create PR linking to issue #$1
4. Provide PR URL

Command Organization

Flat Structure

.claude/commands/
├── create-branch.md
├── review-pr.md
├── run-tests.md
└── deploy.md

Simple, no subdirectories needed.

With Subdirectories

.claude/commands/
├── git/
│   ├── create-branch.md
│   └── create-pr.md
└── testing/
    ├── run-tests.md
    └── coverage.md

Subdirectories organize but don't affect command names:

• /create-branch works regardless of subdirectory
• Description shows "(project:git)" or "(user:testing)"

Best Practices

1. Clear Descriptions

# Good
description: Run all tests and report failures

# Bad
description: Tests

2. Helpful Argument Hints

# Good
argument-hint: [pr-number] [priority] [assignee]

# Bad
argument-hint: [args]

3. Specific Tool Permissions

# Good
allowed-tools: Bash(git status:*), Bash(git add:*), Bash(git commit:*)

# Risky
allowed-tools: Bash(*)

4. Context Before Instructions

# Good
## Current State
!`git status`

## Task
Create commit

# Bad (no context)
Create commit for: $ARGUMENTS

5. Use Heredocs for Commit Messages

# Good
git commit -m "$(cat <<'EOF'
$ARGUMENTS
EOF
)"

# Bad
git commit -m "$ARGUMENTS"  # Formatting issues

Testing Commands

Test Manually

/your-command arg1 arg2

Test Argument Substitution

/review-pr 123 high alice
# Verify $1, $2, $3 substituted correctly

Test Bash Execution

Verify ! commands run and output appears in prompt.

Troubleshooting

Command Not Found

Problem: /mycommand doesn't work

Check:

1. File exists at .claude/commands/mycommand.md
2. Filename matches command (no typos)
3. Restart Claude Code after creating

Arguments Not Substituting

Problem: $1 appears literally in output

Check:

1. Using $1 in prompt content (not frontmatter)
2. Provided arguments when calling command
3. No escaping of $ character

Bash Commands Not Running

Problem: ! commands don't execute

Check:

1. allowed-tools includes Bash permissions
2. Command has ! prefix and backticks: !command``
3. Command syntax is valid

Permission Denied

Problem: Command can't use certain tools

Check:

1. allowed-tools includes needed tools
2. Tool names spelled correctly
3. Bash command patterns match allowed list

Validation Checklist

Before finalizing a command:

• Description is clear and concise
• Argument hints provided (if using args)
• Tool permissions specified (if using tools)
• Arguments use proper format ($ARGUMENTS or $1, $2, etc.)
• Bash commands have ! and backticks
• Bash permissions match commands used
• Heredocs used for multi-line strings
• File tested with sample arguments

Summary

Essential Elements:

1. Clear description (what it does)
2. Argument hints (what to provide)
3. Proper argument substitution ($ARGUMENTS or $1-$9)
4. Tool permissions (if using tools)
5. Context gathering (using ! for bash)

Success Criteria:

• Command shows in /help
• Arguments substitute correctly
• Bash commands execute
• Claude produces expected result
• Tool permissions work