Additional maintainers?

[liamwh]

It seems there's little activity maintainers for the repo, despite there being lots of interest from contributors. How can we get additional maintainers on board so we can move development forward and get the libraries and examples in a good state?

[baxen]

Yes please! Apologies for a few of us initial posters being out last week. We're back, and will make sure to stay active. But we could absolutely use more contributors and would love to combine efforts from other early rust implementations and anyone interested in jumping in

[4t145]

I'd like to help to maintain this crate since I also need to use a rust mcp sdk during my future work in several months.

[jokemanfire]

I also need this crate to develop my project , I want to impove it, By the way, I would like to use it in my project Assistant

[mjio]

This is great. What is the best way for us to combine our efforts? Should we just create PRs? Or should we first come up with ideas about what we think is missing from the Rust SDK and then coordinate our efforts?

[baxen]

Yeah as we get this ready for a more stable release, I think we have two big areas to figure out right now

1. Where do we go with the interface? Some discussion on both Hello developers. I just rewrite this rust SDK with complete and correct data types and do respect the MCP specification. #29 and Defining tools with macros #31 that i think would be good to get a consensus on before we finalize what it looks like to use the crate. Mostly applies to server, but a couple of things on client
2. Implement the missing pieces of the protocol. Some of this also on Hello developers. I just rewrite this rust SDK with complete and correct data types and do respect the MCP specification. #29

I think worth a couple rounds of discussion and then it will be easier to open more concrete items from there?

[jokemanfire]

Is it reasonable ?
#32

[4t145]

Yeah as we get this ready for a more stable release, I think we have two big areas to figure out right now

1. Where do we go with the interface? Some discussion on both Hello developers. I just rewrite this rust SDK with complete and correct data types and do respect the MCP specification. #29 and Defining tools with macros #31 that i think would be good to get a consensus on before we finalize what it looks like to use the crate. Mostly applies to server, but a couple of things on client
2. Implement the missing pieces of the protocol. Some of this also on Hello developers. I just rewrite this rust SDK with complete and correct data types and do respect the MCP specification. #29

I think worth a couple rounds of discussion and then it will be easier to open more concrete items from there?

Maybe we need a place to discuss a blueprint?

[kalvinnchau]

In the interest of getting something concrete to discuss, and with @4t145 putting in a fair amount of work in https://github.com/4t145/rmcp, I'd like to propose:

@4t145 would you be willing to open a PR into this repo with some of your interface changes/updates so that everyone else here can all take a look and review it to see if everyone in the community likes the proposed interface(s)?

Let me know if that sounds like a feasible approach for everyone? We can discuss in PR against the code.

[liamwh]

@mohanz (here), are you interested in contributing here?

[4t145]

In the interest of getting something concrete to discuss, and with @4t145 putting in a fair amount of work in https://github.com/4t145/rmcp, I'd like to propose:

@4t145 would you be willing to open a PR into this repo with some of your interface changes/updates so that everyone else here can all take a look and review it to see if everyone in the community likes the proposed interface(s)?

Let me know if that sounds like a feasible approach for everyone? We can discuss in PR against the code.

Is there any possibility to make rmcp become official sdk? It would be very difficult for me to contribute codes based on current official sdk now.

If there are still some aspects of rmcp that you are not satisfied with, I hope you can point them out. And we can work together to make rmcp meets these requirments. I am confident in my code and I believe it is now better than this library in many aspect.

[baxen]

@4t145 taking a close look at this!

[baxen]

@4t145 I think a good plan is for you to open a PR bringing the rmcp crate into this repo as is, if you are willing to! From there we can make some upgrades before releasing a stable version, and give @cbrit a chance to bring in mcp-types and for us to iterate with @itsaphel on #31 before finalizing.

We'd be excited to add you as a maintainer here. This would however slow down your development a bit, as we'll need to review each change before merging and focus on getting to a stable interface. Does that sound good? I think it will help everyone on this thread have a chance to work on one repo

[4t145]

@baxen Will you merge #44 directly? (I have no idea about "bringing into")