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

# Device Details & Actions

> Open a device to see its system info, authorize it, switch its role, set rules and read its memory, audit its posture, and run response actions on the host.

import { CardGroup, Card, Note, Steps, Step, Tip, Frame } from '@mintlify/components';

Opening a device gives you everything about that one host in a single tab — what it is, whether it's active, how you want the AI to treat it, and the actions you can take on it. Open a device from the **Devices** tab in the [Endpoints view](/documentation/agent-endpoint-security/endpoints-view), or by `@`-mentioning it in chat.

## System Information

The device view opens on its **System Info** — a live snapshot reported by the Myrmex agent:

* **Operating system** and version
* **CPU** and **memory**
* **Agent version** installed on the host
* **Device ID** — the device's unique identifier in Myrmex
* **Public IP** address

This is the same picture [Brontes](/documentation/multi-agent-system/agent-endpoint-manager) keeps current across your whole fleet, focused on a single host.

## Authorizing a Device

A newly installed device first appears as **unauthorized** and does not report until you approve it. This manual step means a device only becomes active in Myrmex once you say so.

<Steps>
  <Step title="Install the agent">
    Onboard the host through the [Download Agent](/documentation/agent-endpoint-security/installation) flow. It registers as **unauthorized**.
  </Step>

  <Step title="Authorize it">
    Open the device and choose **Authorize**. It activates and begins reporting.
  </Step>
</Steps>

You can **Unauthorize** a device later to stop it from reporting without removing it from your inventory.

## Device Role: Agent or Collector

Every device runs in one of two roles, and you can switch between them here:

<CardGroup cols={2}>
  <Card title="Agent" icon="shield-halved" href="/documentation/agent-endpoint-security/endpoint-mode">
    The host protects itself — collecting telemetry and running response actions locally.
  </Card>

  <Card title="Collector" icon="tower-broadcast" href="/documentation/agent-endpoint-security/collector-mode">
    The host acts as a real-time bridge, reaching external systems over SSH or their APIs. Collectors also list their **connected integrations**.
  </Card>
</CardGroup>

## Editing a Device

Adjust how a device is identified and prioritized:

* **Name** and **description** — label the host in terms your team recognizes.
* **Severity** — mark how critical the device is, so investigations and posture reviews weigh it accordingly.
* **Tags** — apply labels that also govern who can access the device.

## Rules and Memory

Two settings shape how the AI works with a device:

<CardGroup cols={2}>
  <Card title="Rule" icon="scroll">
    A free-text, natural-language instruction you write for the AI — for example,
    *"prefer PowerShell on this host"* or *"never restart this server during business hours."* The agents follow it whenever they act on the device.
  </Card>

  <Card title="Memory" icon="brain" href="/documentation/ai-workspace/agent-memory">
    A running record the AI maintains about the device as it learns. It's
    **read-only** — you review it, the AI keeps it current.
  </Card>
</CardGroup>

<Tip>
  Rules also work at the folder level, so you can steer a whole group of devices at once. See [Folders & rules](/documentation/ai-workspace/folders-and-rules).
</Tip>

## Security Posture

A **Security Posture** tab on the device lets you audit that specific host — checking its hardening and configuration against best practice and surfacing findings to fix. See [Security Posture](/documentation/security-posture/overview) for how scoring and findings work across your environment.

## Running Actions on the Host

Response and remediation actions on a device are carried out by [Perseus](/documentation/multi-agent-system/agent-endpoint-specialist), the endpoint specialist, working through the installed agent. Ask in chat — scoping to the device with an `@`-mention — and Perseus can:

* **Isolate the host** from the network to contain an incident.
* **Kill or terminate a process** running on the device.
* **Remediate** — clean up and apply fixes on the host.

<Note>
  Actions that change a host follow the "AI proposes, you approve" pattern — you
  stay in control of what runs on your devices.
</Note>

## Where to Go Next

<CardGroup cols={2}>
  <Card title="Endpoints view" icon="table-list" href="/documentation/agent-endpoint-security/endpoints-view">
    Browse and organize your whole device inventory.
  </Card>

  <Card title="Patch & updates" icon="arrows-rotate" href="/documentation/agent-endpoint-security/patch-management">
    Review and schedule OS updates for this host.
  </Card>

  <Card title="Perseus" icon="screwdriver-wrench" href="/documentation/multi-agent-system/agent-endpoint-specialist">
    The specialist that acts on your hosts.
  </Card>

  <Card title="Endpoint vs Collector" icon="arrows-left-right" href="/documentation/agent-endpoint-security/endpoint-mode">
    How the two device roles differ.
  </Card>
</CardGroup>