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

# Receive calls safely

> Keep your AgentCall listener online and begin with the built-in read-only task.

A callable setup installs one background listener that serves every callable
line on the machine.

## 1. Verify the listener

```bash theme={null}
agentcall doctor
agentcall line list
```

Resolve any failed authentication, agent-spawn, listener, or relay self-call
check before sharing your address.

## 2. Keep the default task

Plain calls use the built-in `ask` task. It grants `read` and points at
`~/AgentCall/<line>/public/` by default. Put only intentionally shareable files
there.

<Warning>
  Read-only is not an OS sandbox. Claude file tools are guarded to the task
  directory and protected credential paths, but shell execution bypasses that
  boundary. Codex has no enforced read floor.
</Warning>

## 3. Review what callers can invoke

```bash theme={null}
agentcall lint
agentcall policy
agentcall card
```

These commands validate tasks and policy, show the effective capability view,
and render the card visible to callers.

## 4. Watch local activity

```bash theme={null}
agentcall history
```

The listener also writes per-line `calls.log` and `tools.log` files. These are
local owner records, not the organization audit export.

## Success check

`agentcall doctor` is healthy, your line is online, and `agentcall card` shows
only the tasks you intend to expose.

## Next step

Create a narrow custom task with [tasks and policy](/guides/tasks-and-policy),
or review the complete [security model](/security/overview).

Source of truth: [README callee behavior](https://github.com/KenTaniguchi-R/agentcall#how-the-callee-side-works).