🧠 Real masters use Claude Code and Codex hidden commands every single day‼️

I completely organized them today (collection-grade)
Claude Code :
1./btw
Perfect for quick side questions. When your main task is running and you suddenly want to ask a small follow-up, use this—won't disrupt your current momentum.
2./rewind
The undo lifesaver. Went off track, broke something, don't like the experiment approach—just roll back and restart. Real efficiency comes from daring to iterate.
3./model opusplan
This one's critical. Strong model handles planning and thinking, fast model handles execution and delivery. Go deep when you need to, move fast when you can.
4./simplify
Code that runs isn't code that's clean. This command is better for cleanup, removing redundancy, and structuring properly.
5./fork
Keep one line of thought, open a new branch to test another approach. Great for architecture work, comparing implementations, or running different paths.
6./remote-control
Perfect for continuing tasks across devices. You don't have to stay glued to your current computer—the session keeps going.
7.Shift + Tab
Easy mode switching. Toggle between execution and planning states without reorganizing your whole workflow each time.
8.Option + P / Alt + P
Quick model switching without clearing your current input. Started typing then want to switch up? Saves tons of time.
Codex:
1./model
Switch models. Use strong model for complex tasks, lightweight model for simple ones—don't run everything through the same setup.
2./permissions
This one matters. Codex can read files, edit files, run commands, so lock down permissions carefully. The more you grant, the more it can do.
3./compact
Essential for long sessions. As context gets longer, compress it—keep what matters, then keep working lighter.
4./fork
Keep current path, create a new branch to test different approaches. Extremely smooth for solution comparison.
5./review
Perfect before committing. Not asking it to rewrite—asking it to code review angle: find risks, spot problems, cover blind spots.
6./diff
See exactly what changed. Don't rely on memory—see the diff directly, most reliable.
7./agent
Great for multi-agent thread scenarios. Different threads handle their tasks, you jump between them and keep pushing forward.
8./fast
Speed up small tasks. When you don't need re-reasoning, just accelerate the pace—feels way more productive.
9.@file
Pull files directly into context. Way more efficient than manually describing file paths.
10.!command
Run local shell directly. For people actually writing code, this capability matters way more than "knowing extra phrases."
11.Enter / Tab / Esc
This shortcut logic is gold. During execution you can still insert commands, queue next steps, restart previous branches—whole interaction feels like you're real-time driving an agent.