> ## Documentation Index
> Fetch the complete documentation index at: https://docs.cogniagent.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Slash commands

> Type / in the chat for the shortcuts that steer your AI coworker — plan, review, keep working, switch model — plus your own skills and workflows.

Type `/` in the message box and a menu appears: the shortcuts that steer your coworker, plus everything you've taught it.

## The commands

| Command    | What it does                                                   |
| ---------- | -------------------------------------------------------------- |
| `/plan`    | Switch to Plan mode — research only, propose a plan            |
| `/build`   | Switch to Build mode — the full toolset                        |
| `/review`  | Turn Thorough review on or off before a task completes         |
| `/loop`    | Keep working on its own. Add a task, or an interval like `10m` |
| `/stop`    | Stop the autonomous loop                                       |
| `/model`   | Show or switch the model for this task                         |
| `/skill`   | Create a skill this coworker can use — optionally name it      |
| `/compact` | Summarise and compact a long conversation                      |
| `/new`     | Start a new task                                               |
| `/clear`   | Clear the message box                                          |
| `/help`    | Show the available commands                                    |

<Frame caption="The command menu. Every entry carries its own one-line description, so there's nothing to memorise.">
  <img src="https://mintcdn.com/glorium/JBtd_griZWPnHu68/images/cowork/chat/slash-commands/01-command-menu.webp?fit=max&auto=format&n=JBtd_griZWPnHu68&q=85&s=03f418452653698c684503a34aa04a34" alt="The slash command menu listing the built-in commands plus a saved skill and workflow" width="864" height="812" data-path="images/cowork/chat/slash-commands/01-command-menu.webp" />
</Frame>

## Your own commands too

The menu isn't only the built-in list. It also holds:

* **Your skills** — each shown as *"Use the **name** skill"*. See [Running a skill](/cowork/skills/running).
* **Your saved workflows** — each shown as *"Run the **name** workflow"*. See [Dynamic workflows](/cowork/capabilities/dynamic-workflows).

Both accept extra text, so `/competitive-analysis Northwind` runs your procedure against that subject.

<Note>
  If a workflow's name clashes with a built-in command or one of your skills, it doesn't get a second row in the menu — one `/deploy` that does two different things would be worse. You can still ask for it in plain words.
</Note>

## The four worth knowing

Most people only ever use a handful.

**`/plan`** before anything expensive or irreversible. Your coworker researches and proposes an approach without changing anything, and you agree it first. See [Plan mode and Build mode](/cowork/chat/plan-and-build).

**`/review`** when being wrong is costly. It checks its own work properly before handing back. See [Thorough review](/cowork/chat/thorough-review).

**`/loop`** for a long job you want finished in one go rather than in check-in-sized pieces. See [Keep working](/cowork/autonomy/keep-working).

**`/compact`** when a conversation has gone on a long time and you'd rather not start a fresh task. See [Long conversations](/cowork/chat/long-conversations).

## Good to know

<Tip>
  You don't have to memorise any of this. Type `/` and read the menu — every command carries its own one-line description.
</Tip>

<Note>
  `/clear` empties the message box you're typing in. It doesn't clear the conversation.
</Note>

## Common questions

<AccordionGroup>
  <Accordion title="Do I need commands at all?" icon="feather">
    No. Plain English covers almost everything — "plan this first" works as well as `/plan`. Commands are quicker once you know them.
  </Accordion>

  <Accordion title="Can I add my own commands?" icon="plus">
    Yes, by [writing a skill](/cowork/skills/creating) or saving a workflow. Both appear in the menu by name.
  </Accordion>

  <Accordion title="What is the difference between /loop and a schedule?" icon="repeat">
    `/loop` keeps working on the task in front of you. A [schedule](/cowork/autonomy/schedules) starts a fresh task on a timetable.
  </Accordion>

  <Accordion title="Does /new lose the current task?" icon="floppy-disk">
    No. It starts a fresh one; the old task stays and can be returned to at any point.
  </Accordion>
</AccordionGroup>

## Next steps

<CardGroup cols={2}>
  <Card title="Plan mode and Build mode" icon="compass-drafting" href="/cowork/chat/plan-and-build">
    The two most useful commands.
  </Card>

  <Card title="Running a skill" icon="play" href="/cowork/skills/running">
    Your own entries in the menu.
  </Card>

  <Card title="Effort and model" icon="gauge-high" href="/cowork/chat/effort-and-model">
    What /model changes.
  </Card>

  <Card title="Keep working" icon="repeat" href="/cowork/autonomy/keep-working">
    What /loop does.
  </Card>
</CardGroup>