# Slash commands and skills

> Run commands and reusable skills from the message field, and control which skills your agents can use.

A skill is a reusable procedure written in a `SKILL.md` file: how to release, how to review a pull request, how to add a component. A command is a saved prompt. Both are one `/` away in the message field.

## Use the / menu

1. Type `/` in the message field. The **Commands and skills** menu opens.
2. Keep typing to filter by name.
3. Use the arrow keys to move, Enter to insert, and Escape to close.
4. Add any details after the name, then send your message.

Each entry shows its description. Entries from the current project carry a **project** badge. Others show their type, such as **skill** or **command**.

## Built-in commands

| Command | What it does |
| --- | --- |
| `/objective <goal>` | Turns on objective mode: your message starts an autonomous loop toward that goal. Available with every provider. |
| `/side <provider> [prompt]` | Opens a parallel conversation with another provider. |
| `/compact` | Summarizes the conversation to free context space. Runs as soon as you pick it. Codex only, once the conversation has started. |
| `/goal <goal>` | Uses the provider's native goal feature. Codex, and Claude when Claude offers it. |

With Claude, the menu also lists the commands that Claude itself provides.

## Where skills come from

Vato gathers skills from these folders. A skill is a folder that contains a `SKILL.md` file.

| Source | On your computer | In a project |
| --- | --- | --- |
| Vato | `~/.vato/skills` | `.vato/skills` |
| Claude | `~/.claude/skills` | `.claude/skills` |
| Codex | `~/.codex/skills` | `.codex/skills` |
| Cursor | `~/.cursor/skills`, `~/.cursor/skills-cursor` | `.cursor/skills`, `.cursor/skills-cursor` |
| Antigravity | `~/.gemini/skills` | `.gemini/skills` |
| Shared | `~/.agents/skills` | `.agents/skills` |

When the same skill exists in several places, a project copy wins over a copy in your home folder. The active provider's own copy wins over Vato's copy.

> [!TIP]
> Put a skill in `~/.vato/skills` to make it available to every provider in Vato.

## Custom commands

A Markdown file in `~/.claude/commands` or in the project's `.claude/commands` folder becomes a command named after the file. For example, `review.md` becomes `/review`. When you send it, the file's content becomes the prompt.

## Manage skills

Open `Settings → Skills` to see every skill found across providers.

- Use the switch next to a skill to enable or disable it. Disabled skills don't appear in agent inputs.
- Use **Restore defaults** to undo your changes.
- **Vato skills folder** shows the shared folder that every provider can use.

### Keep skills in an encrypted cloud vault

Under **Storage**, choose where your personal skills live:

- **Local folders**: `~/.vato/skills` and the provider folders.
- **Encrypted cloud vault**: synced across your devices and encrypted end to end. Vato's server only sees encrypted data.

When you **Create a vault**, Vato shows a recovery code once. Write it down: without it and without another paired device, the vault can't be reopened. To add a device, click **Pair another device** and type the code on the other device within 10 minutes.

### Let agents learn skills

Under **Self-learning**, turn on **Enable skill self-learning**. Agents then regularly check whether finished work deserves a new skill, an update to an existing one, or nothing. Choose where learned skills are saved and how often agents reflect.

Under **Curator**, **Maintain now** reviews skills created by agents. Pinned skills are never changed.

## Next steps

- [Start your first conversation](first-conversation.md)
- [Choose a model](models.md)

Source: https://vatoide.com/docs/agents/slash-commands-and-skills