Skip to content

Conversation

@connor4312
Copy link

Motivation and Context

This is a pattern I've been asked about a few times already, so adding a reference for it.

How Has This Been Tested?

N/A

Breaking Changes

N/A

Types of changes

  • Bug fix (non-breaking change which fixes an issue)
  • New feature (non-breaking change which adds functionality)
  • Breaking change (fix or feature that would cause existing functionality to change)
  • Documentation update

Checklist

  • I have read the MCP Documentation
  • My code follows the repository's style guidelines
  • New and existing tests pass locally
  • I have added appropriate error handling
  • I have added or updated documentation as needed

@jonathanhefner
Copy link
Member

Interesting pattern! Basically, elicitation implemented as an MCP App! Although, for servers using the SHTTP transport, I think it might cause similar reliability issues (around long-running requests, server affinity, etc). What are your thoughts about that?

@connor4312
Copy link
Author

Yup essentially fancy elicitation.

I think the way to tackle reliability is to have the servers set a taskHint for clients that support it 🙂 I'll add mention of this to the example.

@liady
Copy link
Collaborator

liady commented Jan 28, 2026

Nice use case @connor4312 (definitely one that's been asked a lot)!
Agree that the long-running tool issue can be tricky here, and perhaps a task-based implementation will make the pattern too complex?
But a really cool idea, interweaving the two tools like that in the app

@connor4312
Copy link
Author

Maybe, I think right now implementing task support in a server using the TS SDK is a bit complicated. But I know they're still working on it and in theory an SDK (or helper library) could abstract common patterns enough that implementing a task-based tool for a server is not much more expensive than implementing any other kind of tool call. Ideally one would just plug in their choice of persistence layer (memory/redis/etc) into a tool definition and be good to go.

@ochafik ochafik self-requested a review January 29, 2026 17:26
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants