Docker daemon readiness detection, actionable error messaging for DAG resource preparation.

## Problem

When the Docker CLI is installed but the Docker daemon is not running, HomeRail's background DAG resource preparation shows this raw error in the **DAG 资源** status pill:

```text
failed to build homerail-worker:latest (exit code 1)
```

This message is not actionable. It also makes the application look generally broken even though only Docker-backed DAG functionality is unavailable.

A Windows log from the same failure contains the actual cause:

```text
failed to connect to the docker API at npipe:////./pipe/dockerDesktopLinuxEngine
The system cannot find the file specified.
```

## Root cause

The current startup path:

- checks only `docker --version`, which proves that the CLI exists but does not prove that the daemon/server is ready;
- proceeds to image inspection/build and collapses a failed build into the generic `exit code 1` message;
- persists DAG resource state without a structured error code, platform, or recovery action;
- lets the UI display the backend error verbatim.

There is already more suitable readiness logic in `doctor --docker`: it checks the Docker server with `docker version --format "{{.Server.Os}}"` and validates Linux-container mode. The background DAG preparation path should reuse the same logic.

This status path is also separate from the onboarding Docker workspace check, so fixing onboarding alone does not fix the header DAG resource warning.

## Expected behavior

HomeRail's core UI and Agent should remain usable. The warning shoul...