POWER USER FEATURES

Once you're comfortable with basics, these features speed up your workflow significantly.

Built-In Slash Commands

Claude Code has built-in commands starting with /:

/help           Show all available commands
/context        Check context usage
/compact        Compress context manually
/clear          Clear conversation history
/init           Generate CLAUDE.md from codebase
/permissions    View/manage tool permissions

Custom Slash Commands

You can create your own commands. They're just markdown files.

Global commands (work everywhere):

mkdir -p ~/.claude/commands

Project commands (work in specific project):

mkdir -p .claude/commands

Creating a Command

Example: A code review command.

Create ~/.claude/commands/review.md:

Review the code in $ARGUMENTS for:

1. Security vulnerabilities
2. Performance issues  
3. Perl best practices
4. Missing error handling

Be concise and actionable.

Now in any Claude Code session:

/review lib/Auth.pm

The $ARGUMENTS placeholder gets replaced with whatever you type after the command.

Command With Multiple Arguments

Create ~/.claude/commands/compare.md:

Compare $1 and $2:

- What are the key differences?
- Which approach is better and why?
- Any shared patterns?

Use it:

/compare lib/OldAuth.pm lib/NewAuth.pm

$1 is first argument, $2 is second, etc.

Useful Custom Commands

Safe commit checker:

~/.claude/commands/commit.md:

Review all uncommitted changes (git diff HEAD).

Before I commit, check for:
- TODO or FIXME comments that shouldn't be committed
- Commented-out code
- Debug statements (print, console.log, dd())
- Hardcoded test values
- Sensitive data (API keys, passwords)

If clean, suggest a commit message.
If issues found, list them.

Test writer:

~/.claude/commands/test.md:

Write tests for $ARGUMENTS:

- Test normal operation
- Test edge cases
- Test error conditions
- Use Test::More conventions
- Include setup/teardown if needed

Explain command:

~/.claude/commands/explain.md:

Explain $ARGUMENTS in detail:

- What does it do?
- How does it work?
- What are the key parts?
- Any non-obvious behavior?

Assume I'm familiar with Perl but new to this codebase.

Session Management

Continue last session:

claude --continue

Resume a specific session:

claude --resume

This opens a picker showing recent sessions.

Useful when:

• Terminal crashed
• You closed it accidentally
• Want to revisit yesterday's work

Context Commands

Check usage:

/context

Shows:

• Current context percentage
• Which files are loaded
• CLAUDE.md status

Manual compact:

/compact

Summarizes conversation and clears old messages. Good habit after completing a major task.

Auto-compact threshold (in settings.json):

"env": {
  "CLAUDE_AUTOCOMPACT_PCT_OVERRIDE": "75"
}

Triggers automatic compaction at 75% context usage.

File Commands

Add additional directories:

/add-dir ~/shared/libraries

Now Claude can see files outside your current project. Useful for referencing shared code.

Permission Management

Interactive permission editor:

/permissions

Shows what's allowed/denied. You can adjust here instead of editing settings.json.

Status Line

Claude Code shows status info at the bottom. Customize with:

/statusline

Shows model, context usage, and other info. Useful for monitoring.

MCP Integration

Model Context Protocol (MCP) lets you add external tools. Not covered in depth here, but you can add MCP servers:

claude mcp add postgres

Then Claude can query databases, interact with APIs, etc.

View configured MCP servers:

/mcp

Working on Multiple Files

If you're changing many related files:

> Read @lib/Controller/*.pm and understand the pattern.
> Now apply the same pattern to @lib/Controller/NewThing.pm

Claude keeps the pattern consistent across files.

Batch Operations

For repetitive tasks across many files:

> For each file in @lib/Model/*.pm:
> 1. Add proper documentation header
> 2. Add "use strict; use warnings;"
> 3. List what you changed

Claude processes them systematically.

Debugging Sessions

When Claude seems stuck or confused:

/context

Check if context is nearly full. If yes, /compact.

> Summarize what we've discussed so far

See if Claude's understanding matches yours.

> Start fresh on this problem

Reset approach without clearing history.

Tips From Power Users

1. Keep commands short and focused

A 500-word command file costs context every time it loads.

2. Use project commands for project-specific tasks

Global: generic utilities
Project: specific to that codebase

3. Chain commands mentally

/review lib/App.pm
[fix issues]
/test lib/App.pm
[verify]
/commit

4. Alias claude for convenience

Add to ~/.zshrc:

alias c='claude'
alias cc='claude --continue'

5. Use /compact after major milestones

Finished a feature? /compact
Fixed a bug? /compact
Don't wait for context to fill

Summary

Custom commands:
  ~/.claude/commands/name.md     Global
  .claude/commands/name.md       Project
  Use $ARGUMENTS, $1, $2, etc.

Session management:
  claude --continue              Resume last
  claude --resume                Pick from history

Context management:
  /context                       Check usage
  /compact                       Compress
  Auto-compact via settings.json

Built-in commands:
  /help, /init, /permissions, /mcp

Next chapter: Troubleshooting common issues.