Claude Code

Guide to working effectively with Claude Code CLI.

Purpose

Provide quick reference for:

• Essential commands and shortcuts
• Configuration and settings
• Best practices for effective prompting
• Plugin and skill development
• Workflow optimization

Essential Commands

Slash Commands

Keyboard Shortcuts

Context Management

CLAUDE.md Files

Project instructions stored in CLAUDE.md:

markdown
1# Project Instructions
2
3## Build Commands
4- `npm run build` - Build project
5- `npm test` - Run tests
6
7## Code Style
8- Use TypeScript strict mode
9- Prefer functional components
10
11## Important Files
12- `src/config.ts` - Configuration
13- `src/types/` - Type definitions

Location Priority:

1. ./CLAUDE.md - Current directory (highest)
2. ~/.claude/CLAUDE.md - Global (lowest)

Memory Files

Persistent memory across sessions:

bash
1# View memory
2/memory
3
4# Memory stored in
5~/.claude/memory.md

Tool Permissions

Permission Levels

Configure Permissions

bash
1/permissions
2
3# Or in settings
4/config

Common Permission Patterns

yaml
1# Allow running tests
2- tool: Bash
3  prompt: "run tests"
4
5# Allow npm commands
6- tool: Bash
7  prompt: "npm"
8
9# Allow git operations
10- tool: Bash
11  prompt: "git"

Effective Prompting

Be Specific

# Good
"Add error handling to the fetchUser function in src/api/users.ts
that catches network errors and returns a default user object"

# Less effective
"Add error handling to the code"

Provide Context

# Good
"The build is failing with 'Cannot find module ./utils'.
Check the import paths in src/components/Header.tsx"

# Less effective
"Fix the build error"

Break Down Complex Tasks

# Step by step
1. "First, read the current implementation in src/auth/"
2. "Now add a logout function that clears the session"
3. "Update the tests in tests/auth.test.ts"

Use File References

# Reference specific files
"Update @src/config.ts to add a new API endpoint"

# Reference multiple files
"Compare @src/old.ts with @src/new.ts and list differences"

Model Selection

Available Models

Switch Models

bash
1/model opus    # For complex tasks
2/model sonnet  # Default balance
3/model haiku   # Quick operations

Working with Code

Read Before Edit

Always read files before modifying:

# Claude will read first
"Update the validation in src/forms/login.ts"

# Explicit read
"Read src/forms/login.ts and explain the validation logic"

Incremental Changes

# Prefer small, focused changes
"Add input validation to the email field"

# Then
"Add input validation to the password field"

# Rather than
"Rewrite the entire form with validation"

Test After Changes

"Run the tests to verify the changes work"

# Or be specific
"Run npm test -- --grep 'login'"

Git Integration

Common Workflows

# Check status
"Show git status"

# Create commit
"Commit these changes with message: Add user validation"

# Create PR
"Create a pull request for this feature"

# Review PR
/review

Commit Best Practices

Claude auto-adds co-author:

Co-Authored-By: Claude <noreply@anthropic.com>

Background Tasks

Run Tasks in Background

# Long-running commands
"Run the build in the background"

# Check status
/tasks

# View output
"Show output from task abc123"

Plugins & Skills

Available Skills

Invoke with slash commands:

/commit          # Create git commit
/review          # Review code changes
/pr-comments     # View PR feedback

Custom Skills

Create in .claude/skills/ or .windsurf/skills/:

skills/
└── my-skill/
    ├── SKILL.md         # Skill definition
    └── references/      # Supporting docs

See references/skill-development.md for details.

Troubleshooting

Common Issues

Debug Mode

bash
1# Verbose output
2claude --debug
3
4# Check configuration
5claude --version
6/config

Reset Session

bash
1# Clear conversation
2/clear
3
4# Fresh start
5# Exit and restart claude

Best Practices Summary

1. Read before writing - Understand existing code first
2. Be specific - Clear instructions get better results
3. Incremental changes - Small steps, verify each
4. Use context files - Maintain CLAUDE.md for projects
5. Manage permissions - Set up trusted patterns
6. Use the right model - Opus for complex, Haiku for simple
7. Compact regularly - Keep context manageable

Additional Resources

Reference Files

• references/skill-development.md - Creating custom skills
• references/hooks.md - Event hooks for automation

External Resources

• GitHub Issues: https://github.com/anthropics/claude-code/issues
• Documentation: /help command