Commands Overview

ClaudeKit provides a comprehensive set of slash commands to accelerate your development workflow. Each command is designed for specific tasks and automatically orchestrates the appropriate agents.

Command Categories

Core Development

/bootstrap - Initialize new projects with spec-driven development
/ask - Ask questions about the codebase
/ck-help - Get help about ClaudeKit commands
/coding-level - Set coding complexity level
/cook - Implement features with intelligent workflow routing
/cook
- Automatic feature implementation
/cook:auto
- Quick automatic implementation
/cook:auto
- Parallel automatic implementation
/debug - Investigate and diagnose issues
/journal - Write development journal entries
/kanban - Manage project tasks in Kanban board
/preview - Preview current work
/test - Run test suite and get results
/use-mcp - Use MCP server tools
/watzup - Get project status and recent changes
/worktree - Manage git worktrees
/wt - Shorthand for /worktree

Planning

/plan - Create implementation plans
/plan
- Quick planning
/plan
- Detailed planning with research
/plan
- Create plan with 2 approaches
/plan
- Parallel planning with multiple researchers
/plan
- Analyze CI failures and create fix plan
/plan
- Create conversion optimization plan
/plan
- Archive completed plans
/plan
- Validate plan structure

Documentation

/docs - Manage project documentation
/docs
- Initialize project documentation
/docs
- Update project documentation
/docs
- Summarize project documentation

Review

/review
- Review entire codebase
/review:codebase
- Parallel codebase review

Other

/check-and-commit - Check code quality and commit

Quick Command Reference

Most Used Commands

# Project Setup
/bootstrap [description]         # Initialize new project
/coding-level [level]            # Set coding complexity level

# Feature Development
/plan [feature description]      # Plan the feature
/plan:hard [complex feature]     # Detailed planning with research
/plan:parallel [feature]         # Parallel planning

# Documentation
/docs:init                       # First-time setup
/docs:update                     # After making changes
/docs:summarize                  # Generate documentation summary

# Debugging & Testing
/debug [issue]                   # Investigate and diagnose
/test                            # Run test suite
/test:ui                         # Run UI tests

# Git Workflow
/check-and-commit                # Check quality and commit
/worktree [feature-name]         # Create git worktree

# Project Status
/watzup                          # What's the current state?
/ask [question]                  # Ask about codebase
/kanban                          # View project dashboard

Command Syntax

Basic Syntax

/command [required-argument] [optional-argument]

Examples

# No arguments
/test
/watzup
/docs:init

# Required argument
/cook [add user authentication]
/debug [login button not working]
/ask [how does routing work?]

# Optional arguments
/git:pr                          # PR to default branch
/git:pr [develop]                # PR to develop
/git:pr [main] [feature-branch]  # PR from feature to main

# Multiple arguments
/scout [authentication files] [3]  # Find auth files, use 3 agents

Command Workflows

Starting a New Project

1. /bootstrap [project description]
   # OR
   ck init --kit engineer

2. # Customize requirements through Q&A

3. # System automatically:
   - Researches best practices
   - Creates implementation plan
   - Implements features
   - Generates tests
   - Sets up documentation

Developing a Feature

1. /plan [feature description]
   # Creates detailed implementation plan

2. # Review plan in plans/ directory

3. # Implement based on plan
   # Work with fullstack-developer agent

4. /test
   # Validates implementation

5. /check-and-commit
   # Checks quality and commits

Investigating Issues

# Debug issues
/debug [describe issue]
# - Analyzes logs and code
# - Identifies root cause
# - Suggests solutions

# Run tests
/test
# - Runs full test suite
# - Shows coverage report

# UI testing
/test:ui
# - Runs UI/E2E tests
# - Visual regression testing

Updating Documentation

# After implementing features
/docs:update

# When onboarding new team members
/docs:summarize

# When starting with existing codebase
/docs:init

Command Best Practices

Use the Right Command for the Task

✅ Correct Usage

# Simple questions
/ask [how does routing work?]

# Planning with research
/plan:hard [add OAuth2 authentication]

# Quick planning
/plan:fast [update button styling]

# Parallel planning for complex features
/plan:parallel [rebuild authentication system]

❌ Incorrect Usage

# Don't use fast for complex features
/plan:fast [rebuild entire authentication system]

# Don't skip planning for major features
# (implement without /plan first)

Provide Clear Descriptions

✅ Clear

/plan [add OAuth2 authentication with Google and GitHub providers]
/debug [API returns 500 error when creating user with empty email]
/ask [how is user authentication currently implemented?]

❌ Vague

/plan [add auth]
/debug [something's broken]
/ask [how does this work?]

Review Before Committing

# 1. Implement feature

# 2. Test
/test

# 3. Review changes
git diff

# 4. Quality check and commit
/check-and-commit

Use Sequential Commands for Complex Tasks

# 1. Understand codebase
/ask [how is authentication currently implemented?]

# 2. Plan changes with research
/plan:hard [migrate from session-based to JWT authentication]

# 3. Review plan
/preview plans/latest-plan.md

# 4. Implement based on plan
# Work with agents to implement

# 5. Test
/test

# 6. Debug if needed
/debug [specific issue]

# 7. Quality check and commit
/check-and-commit

Command Flags and Options

Some commands support flags:

/bootstrap

/bootstrap [project description]              # Interactive Q&A
/bootstrap:auto [detailed description]        # Fully automatic
/bootstrap:auto:fast [description]            # Quick automatic
/bootstrap:auto:parallel [description]        # Parallel execution

/plan

/plan [feature]              # Intelligent planning
/plan:fast [feature]         # Quick planning without research
/plan:hard [feature]         # Detailed planning with research
/plan:two [feature]          # Two different approaches
/plan:parallel [feature]     # Parallel planning with researchers
/plan:ci [CI-URL]            # Analyze CI failures
/plan:cro [content]          # Conversion optimization

Understanding Command Output

Commands provide structured output:

Planning Commands

planner Agent: Analyzing codebase...

Research Results:
- OAuth2 best practices reviewed
- Existing auth patterns identified
- Security considerations documented

Implementation Plan Created:
📄 plans/oauth-implementation.md

Plan Summary:
1. Install dependencies (passport, passport-google-oauth20)
2. Configure OAuth2 providers
3. Implement callback routes
4. Add session management
5. Generate tests
6. Update documentation

Estimated time: 2-3 hours
Files to create: 5
Files to modify: 3

Next: Review plan, then implement

Implementation Output

fullstack-developer Agent: Implementing from plan...

Dependencies Installed:
✓ passport (0.6.0)
✓ passport-google-oauth20 (2.0.0)

Files Created:
✓ src/auth/oauth-config.js
✓ src/auth/google-strategy.js
✓ src/routes/auth-callback.js

Tests Generated:
✓ tests/auth/oauth.test.js (15 tests)

Documentation Updated:
✓ docs/api/authentication.md

Implementation complete!

Next: Run /test to validate

Test Commands

tester Agent: Running test suite...

Test Results:
✓ Unit tests: 45 passed
✓ Integration tests: 12 passed
✓ E2E tests: 8 passed

Coverage: 87.3%

All tests passed!

Next: Review changes, then /git:cm

Troubleshooting Commands

Command Not Found

Problem: /command not recognized

Solutions:

Verify you’re in a ClaudeKit project (ls .claude/)
Check command exists (ls .claude/commands/)
Run ck init to get latest commands
Restart Claude Code

Command Fails

Problem: Command errors during execution

Solutions:

Check error message for specific issue
Verify prerequisites (API keys, dependencies)
Review agent logs
Try command with simpler input
Use /debug to investigate

Unexpected Results

Problem: Command doesn’t do what expected

Solutions:

Review command documentation
Check if using correct command for task
Provide more specific description
Review generated plans before implementing
Use feedback to refine

Next Steps

Explore specific command categories:

Core Commands - Development essentials
Planning Commands - Implementation planning
Documentation Commands - Documentation management
Review Commands - Code review and analysis

Or learn about:

Agents - How commands invoke agents
Skills - Specialized capabilities
Quick Start - Hands-on tutorial

Key Takeaway: ClaudeKit commands provide a natural, intuitive interface to powerful agent orchestration, making complex development tasks simple and repeatable.