upvote

points

by acchow2 hours ago |
comments
upvoteby Terretta10 minutes ago|
next
[-]
> > Every MCP server injects its full tool schemas into context on every turn

> I consider this a bug. I'm sure the chat clients will fix this soon enough.

ANTHROP\C's Claudes manage/minimize/mitigate this reaonably.

reply
upvoteby phh2 hours ago|
prev|
next
[-]
Absoultely.

I'll add to your comment that it isn't a bug of MCP itself. MCP doesn't specify what the LLM sees. It's a bug of the MCP client.

In my toy chatbot, I implement MCP as pseudo-python for the LLM, dropping typing info, and giving the tool infos as abruptly as possible, just a line - function_name(mandatory arg1 name, mandatory arg2 name): Description

(I don't recommend doing that, it's largely obsolete, my point is simply that you feed the LLM whatever you want, MCP doesn't mandate anything. tbh it doesn't even mandate that it feeds into a LLM, hence the MCP CLIs)

reply
upvoteby fennecbutt2 hours ago|
prev|
next
[-]
Yup, routing is key. Just like how we've had RAG so we don't have to add every biz doc to the context.

I agree with the general idea that models are better trained to use popular cli tools like directory navigation etc, but outside of ls and ps etc the difference isn't really there, new clis are just as confusing to the model as new mcps.

reply
upvoteby edgyquant1 hours ago|
prev|
next
[-]
That’s a trade off, now you need multiple model calls for every single request
reply
upvoteby ekianjo2 hours ago|
prev|
[-]
Yes we just RAG to be applied on tools. Very simple to implement.
reply
upvoteby edgyquant1 hours ago|
parent|
[-]
I don’t think so. Without a list of tools in context the ai can’t even know what options it has, so a RAG like search doesn’t feel like it would be anywhere near as accurate
reply
upvoteby ekianjo14 minutes ago|
parent|
[-]
The RAG helps select the tool needed for the task at hand. Semantic search returns only the tools that match. Very efficient.
reply