Auto-install runner backend when running a model

[doringeman]

When a user runs docker model run , if the required backend (e.g. vllm) is not installed, the CLI currently reports it as "not installed" with no further action. The user must separately run docker model install-runner --backend vllm before their original command will work.

Current behavior:

• docker model status reports vllm as "not installed"
• User must manually run docker model install-runner --backend vllm
• User re-runs their original docker model run command

Desired behavior:
When a user runs docker model run <model> and the required backend is not installed, the CLI should:

• Detect that the backend is missing
• Prompt the user to confirm that the backend can be downloaded and used
• Download and install the backend automatically upon confirmation
• Proceed with running the model

This eliminates the need for an explicit, separate install command while still giving the user visibility and control over what's being downloaded.