refactor(cli): extract service settings and inference commands

[varshaprasad96]

Summary

Extract the service, settings, and inference command implementations from the CLI run module into focused command modules. Preserve the existing run-module re-exports so command dispatch and downstream callers remain compatible.

Related Issue

Part of #2304

Changes

• Add dedicated service, settings, and inference command modules.
• Move existing implementations and tests to their owning modules.
• Keep shared setting parsing tests with the common command helpers.
• Add focused coverage for TCP forward parsing, settings JSON rendering, and inference status errors.

Testing

• mise run pre-commit passes
• Unit tests added/updated
• E2E tests not applicable; no sandbox infrastructure, policy, or user-visible behavior changed
• mise run test passes
• mise run ci passes

Checklist

• Follows Conventional Commits
• Commits are signed off (DCO)
• Architecture docs not applicable; module ownership changed without an architecture or behavior change

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[johntmyers]

not sure we need to extract inference, see #3172

[varshaprasad96]

@johntmyers thanks for pointing this out! Just looked into the other issue, this one needs to be modified then to refactor only service settings, and remove inference. I'm going to point the agent to handle both of it in one go and modify this PR
/hold

[varshaprasad96]

Closing this in favor of #3194