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

# Command execution

> How MCP command strings route and how agents should call them.

`aeo_execute_command` accepts the command after the `aeo` prefix.

```json theme={null}
{
  "command": "visibility show",
  "domainId": "..."
}
```

Not:

```json theme={null}
{
  "command": "aeo visibility show"
}
```

## Execution examples

### Agent context

```json theme={null}
{
  "name": "aeo_execute_command",
  "arguments": {
    "command": "agent context",
    "domainId": "<domain_id>"
  }
}
```

### Import a draft

MCP clients cannot send local `--body-file` paths. Pass inline body text through the command registry when importing from a remote agent.

This is a write operation. Use a `read_write` token and only run it after the user has approved importing the draft.

```json theme={null}
{
  "name": "aeo_execute_command",
  "arguments": {
    "command": "content import --title \"Runner sunscreen guide\" --body \"# Runner sunscreen guide\\n...\"",
    "domainId": "..."
  }
}
```

### Read a reference style

```json theme={null}
{
  "name": "aeo_execute_command",
  "arguments": {
    "command": "reference style --url https://www.instagram.com/p/...",
    "domainId": "..."
  }
}
```

## Agent safeguards

* Read `agent context` before writing.
* Confirm before credit-metered jobs.
* Confirm before public preview links.
* Confirm before deploy or publish.
* Use `read` tokens for discovery; use `read_write` only for an approved write.
* Prefer import/update workflows over server-side generation unless the user explicitly asks for generation.