5 Underrated Claude Code Commands for Everyday Use
If you use Claude Code and you're still living in the chat box, you're working twice as hard as you need to. There's a whole command layer sitting underneath it, and most people never touch it. I use five slash commands every single day. They cut repetition, automate the boring stuff, and let Claude finish long tasks without me babysitting. The last one has completely changed how I work.
Why this matters
Most people treat Claude Code like a chat window. You type, it responds, you type again. But Claude Code also has slash commands: one-word shortcuts that trigger whole workflows. You type a forward slash, then the command name, and Claude handles the rest. Once you start using them, going back to plain chat feels like doing everything with a trackpad when you have keyboard shortcuts.
Key insight: Slash commands aren't power-user tricks. They're the fastest way to stop re-explaining yourself and start getting work done on autopilot.
Commands I use every day
1. /memory to stop re-explaining yourself
Type /memory and Claude updates your CLAUDE.md and context files. That's Claude's memory across sessions. You write in your preferences, your rules, your facts about how you work. Claude reads those files every time it starts up.
Without this, you're starting from scratch every session. You re-explain your tone, your stack, your team, your pet peeves. With it, Claude already knows. I treat /memory like a quick save button for anything I want to stick.
2. /schedule to get work delivered on autopilot
/schedule runs a prompt on a timer. You set it once and forget it. I have mine send me a daily brief every morning at 9am. I don't open Claude, don't type anything, don't check in. It just shows up.
This is great for recurring tasks: morning summaries, weekly check-ins, reminders to review a dashboard. Anything you'd otherwise forget to ask for.
3. /loop to keep a check running while you do something else
/loop keeps a prompt running on repeat while your session is open. Think of it as a background monitor. Claude runs the same check over and over until you stop it.
I use this when I'm waiting on a process to finish, watching a deploy, or checking whether something has changed. I can work on something else while Claude keeps an eye on it.
4. /rewind when you need to undo, not explain
/rewind rolls your conversation and code back to an earlier checkpoint. If you make a mistake, don't type "revert back to the previous version." Claude will get confused trying to interpret that as a normal request.
Just use /rewind. It snaps you back to a clean point in your session. Way faster than untangling a mess you made three prompts ago.
5. /goal to let Claude work until it's actually done
/goal is my favorite right now. You set a condition, like "all tests passing" or "build completes without errors," and Claude works on its own until it gets there. You don't have to keep nudging it. You don't have to say "keep going" every few minutes.
This is perfect for long-running tasks you want finished through completion. Hand Claude the goal, walk away, come back when it's done.
Here are some related guides to check out: