Hidden and Underused Features in Claude Code

1. Mobile App for Coding Anywhere

Claude Code is not limited to desktop.

You can write and modify code directly from the iOS or Android app. This is surprisingly practical for quick fixes, reviewing changes, or iterating on ideas without opening your laptop.

2. Seamless Session Handoff Across Devices

You can move sessions between mobile, web, desktop, and terminal.

• claude --teleport or /teleport — Continue a cloud session locally
• /remote-control — Control a local session from your phone or browser

This turns Claude into a continuous workspace rather than a device-bound tool.

3. Automation with /loop and /schedule

Two of the most powerful features:

• /loop — Let Claude run tasks repeatedly
• /schedule — Execute at intervals for up to a week

Example use cases:

• Auto-fixing code issues every 5 minutes
• Running continuous code review
• Keeping branches rebased

This is where Claude starts behaving like an autonomous agent.

4. Hooks for Deterministic Workflows

Hooks let you inject logic into the agent lifecycle.

You can:

• Load context automatically on session start
• Log every command Claude runs
• Route permission requests to external tools like WhatsApp

This gives you fine-grained control over behavior and observability.

5. Cowork Dispatch for Remote Execution

Dispatch allows you to control your Claude Desktop environment remotely.

You can:

• Manage files
• Handle Slack and email
• Execute tasks from your phone

It effectively turns Claude into a remote operating layer for your machine.

6. Chrome Extension for Frontend Work

Claude can connect directly to your browser.

This enables:

• Debugging with console logs
• Testing UI changes
• Automating browser workflows

The key insight is simple: give Claude a way to verify its output, and it will iterate until the result is correct.

7. Auto Running and Testing Web Servers

Claude Desktop can automatically:

• Start your web server
• Test it in a built-in browser

This removes a lot of manual steps in development loops.

8. Forking Sessions

You can branch your work like Git.

Two ways:

1. /branch inside a session
2. claude --resume <session-id> --fork-session from CLI

This is extremely useful for exploring different approaches without losing progress.

9. /btw for Side Queries

Use /btw to ask quick questions while your main task is running.

Example:

• Clarify syntax
• Check definitions
• Ask unrelated questions

This keeps your workflow uninterrupted.

10. Git Worktrees for Parallel Work

Claude has deep support for git worktrees.

This allows:

• Multiple parallel environments
• Running many agents on the same repo
• True parallel development

Essential if you are scaling agent-based workflows.

11. /batch for Massive Parallel Changes

/batch lets Claude distribute work across many agents.

It:

• Interviews you about the task
• Splits work into multiple worktrees
• Executes in parallel across dozens or hundreds of agents

Perfect for migrations and large-scale refactoring.

12. --bare for Faster SDK Execution

By default, Claude loads local configs, MCPs, and settings.

With --bare, you:

• Skip unnecessary loading
• Speed up execution
• Gain more control in scripted workflows

13. --add-dir for Multi-Repo Context

When working across repositories:

Use --add-dir to give Claude access to additional folders.

This allows:

• Cross-repo understanding
• Shared context
• Safer multi-project operations

14. Custom Agents with --agent

You can define your own agents with custom prompts and tools.

Steps:

1. Create an agent in .claude/agents
2. Run with claude --agent=<name>

This turns Claude into a system of specialized agents rather than a single general model.

15. Voice Input for Coding

You can code by speaking.

Use:

• /voice in CLI
• Voice button in Desktop
• iOS dictation

Speaking often leads to more detailed instructions than typing, which improves output quality.