Plan an MCP Server's Tools and Resources

I want to build an MCP server for this use case:

{{use_case}}

Systems/data it needs to connect to: {{systems}}

Design the server as:
1. A list of tools (actions the model can take), each with a name, a one-line
   description written for the calling model, and its key parameters.
2. A list of resources (read-only context the model can pull in), if any make sense
   for this use case — and an explicit note if none do.
3. For each tool, whether it's read-only or has side effects, and if it has side
   effects, what confirmation or dry-run mode it should support.
4. Any place where a single tool is trying to do two things and should be split, or
   two tools are redundant and should be merged.

Optimize for a small number of clearly-scoped tools over broad coverage — flag if my
use case is trying to cram too much into one server and would be better split into two.