Command
A reusable instruction you run by typing /name. See what it is, how it works, how it is used and what to watch for, with pictures and examples.
What it is
A command is a reusable instruction you run by typing a slash and a name, like /clear. Built-in commands control the session, for example switching , clearing or checking what fills the . You can also make your own commands as files or get them from a plugin. The old name "slash command" means the same thing.
How it works
-
Type / at the
-
Keep typing to filter the list
-
Pick a command and press Enter
-
Text after the name becomes arguments
How to use it
-
/clearstarts fresh on a new task while keeping project .
-
/modeland/effortchange the model and how much reasoning it uses.
-
/contextand/compactsee what fills the context window, then summarize to free space.
-
.claude/commands/deploy.mda file like this creates your own
/deploycommand.
Watch out
A command is only recognized at the start of your message. For new work, the docs recommend a in .claude/skills/ instead of a file in .claude/commands/, because skills can hold extra files and can load them by itself. Do not confuse it with such as claude mcp add, which you type in the , not inside a session.
Related terms
Original and official docs
This page explains the term in plain words. The original is the authority: if they differ, follow the original.
- See this term in the originalClaude Code Docs — Glossary Anthropic · checked 2026-10-19
- Anthropic — Claude Code docs: Commands — All commands(opens in a new window)
- Anthropic — Claude Code docs: Extend Claude with skills — How a skill gets its command name(opens in a new window)
- Anthropic — Commands(opens in a new window)