Reference

Slash commands

Every slash command in the LoopCode terminal, grouped by what it is for, plus the commands available in the desktop app and VS Code.

Type / in the terminal to see the commands, filtered as you type. Your skills and commands appear in the same list, and run with /name arguments.

Mode and model

Command	Does
`/model`	Choose the model
`/effort`	Choose how hard the model works, or turn adaptive effort on or off (`/effort adaptive on`)
`/mode`	Choose what LoopCode is allowed to do
`/plan`	Switch to Plan mode
`/thinking`	Show or hide the model's reasoning

Command	Does
`/init`	Create (or improve) a `LOOPCODE.md` with instructions for this project
`/review`	Review your current changes and find issues. Changes no files
`/diff`	Show the git diff, including untracked files
`/pr [title]`	Commit, push and open a GitHub pull request with the current work
`/agents`	List the subagents LoopCode can delegate to
`/goal <objective>`	Set a goal: LoopCode keeps working until it can prove it done. `/goal` alone shows and steers it
`/loop [interval] <prompt>`	Repeat a prompt on an interval, or let LoopCode pace itself
`/automations`	Manage scheduled jobs
`/worktree`	Start or continue a conversation in a new worktree
`/tasks`	List background commands, monitors and subagents; show output; stop them
`/stop`	Stop all background tasks

Command	Does
`/new`	Start a new conversation
`/resume`	Resume a saved conversation
`/rename`	Rename the current conversation
`/fork`	Continue in a copy of this conversation, to try another way; the original stays
`/rewind`	Undo the file changes of recent turns
`/compact [what to keep]`	Summarise the conversation to free context
`/btw <question>`	Ask something on the side; the answer is not added to the conversation
`/recap`	Recap the conversation so far, without adding to it
`/mention`	Mention a file
`/attach [path]`	Attach a file (PDF, document, code) or an image to the next message
`/copy`	Copy the last response
`/export`	Export the conversation as Markdown
`/remote`	Control this session from your phone with a QR code; `/remote` again turns it off
`/archive`	Archive this conversation and start a new one
`/delete`	Delete this conversation for good

Command	Does
`/memory`	See and edit what LoopCode remembers and has learned
`/skills`	Use and manage skills, including learned ones
`/hooks`	View the lifecycle hooks in force
`/mcp`	Manage MCP servers
`/plugins`	Browse plugins
`/import`	Bring in the MCP servers you set up for Claude Code or Codex
`/lsp`	Show the language servers behind code intelligence
`/permissions`	Show the permission rules in force
`/status`	Show the session's configuration and token usage
`/context`	Show what fills the context window
`/cd <dir>`	Work in another folder from here on (a fresh conversation there)
`/pwd`	Show the folder LoopCode works in

Command	Does
`/usage`	View plan usage and limits
`/upgrade`	Upgrade or renew Pro or Max
`/login`	Sign in to your LoopCode plan
`/logout`	Sign out of LoopCode
`/feedback [message]`	Write to the LoopCode team (opens your mail app)

Command	Does
`/help`	Show commands and keys
`/clear`	Clear the terminal and start a new conversation
`/exit`	Exit LoopCode

Commands you may know from other tools work too:

The composer in the desktop app and VS Code offers your skills plus:

Command	Does
`/loop [interval] <prompt>`	Repeat a prompt on a schedule, or let the agent pace itself
`/compact [what to keep]`	Summarise the conversation now, to free context
`/context`	See what fills the context window
`/btw <question>`	Ask on the side; the answer is not added to the conversation
`/recap`	Recap the conversation so far

The rest — models, modes, memory, connections, skills, automations — is in the app's pickers, pages and command palette (⌘ K). See Desktop app.