CLI missing many features for server management

Nobody has claimed this yet.

I'm attempting to use a headless server with a K80 as my main AI box, but this doesn't really seem possible at the moment with lms and llmster. lms is missing many of the features of the desktop app, or is missing documentation for such features, which basically means they don't exist. The only options for `lms server` are start, stop, and status. Things that I can do on desktop but seemingly not with lms:

1. set the server port number
2. manage api keys
3. enable/disable authentication
4. enable/disable serving on the local network, which seems to be necessary for other PC's to be able to use the API
5. manage mcp servers (I did find the mcp.json file, so seemingly that exists but I didn't see documented anywhere)
6. enable/disable api calls to said MCP servers
7. enable/disable JIT model loading
8. set the TTL for JIT model loading
9. enable/disable CORS

Is there some config file somewhere I'm missing?

thanks!

Start by reviewing the documented `lms server` entry point and the existing `mcp.json` configuration. Compare the nine desktop capabilities listed in the issue with what the CLI documentation covers, and clarify which are supported, configurable, or absent. Done means the available server-management options and configuration limits are documented clearly.

Written by the indexing model from the issue text.