lets get to the best version faster - follow @1dolinski · challenge your habits · talk to my AI (in this frame https://warpcast.com/1dolinski/0xc35ca78f)

While building skeet.build I’m quickly finding out a few thing:

1. Community mcp servers are half broken all the time and don’t really work
2. People think mcp servers are just wrappers around APIs, so it’s designed poorly. The best example is the linear mcp:

To move an issue from todo to in progress, the api requires a uuid of all the different states. If you plugin an open api v3 spec then your llm will be like I’m sorry don’t have the proper inputs and just hang. Or you have to make like 3 other api calls to do one task.

It actually requires thought to make the experience just “work”

I thought this was an interesting point by @jj https://warpcast.com/jj/0xb56d0d0a and decided to explore it

With just my Linear personal API key and no docs, I built a Nash Task that can:
- list issues
- mark issues as done
- create issues
- move issues to a different state
- list workflow states (I don't even know what this is)

Tasks are shareable. Now I can give this to my team or anyone. No extra MCP to install. Maybe your team's linear workflow is slightly different. That's fine, just have Nash update the task to meet your needs.

With the rapid progress of these models, I just don't buy the idea we'll have thousands of one-off MCPs to interact with services.

🗽based; 

building skeet.build 🟠;
building @buoy 🛟; 
building check.supply 🏦;

I thought this was an interesting point by @jj https://warpcast.com/jj/0xb56d0d0a and decided to explore it

With just my Linear personal API key and no docs, I built a Nash Task that can:
- list issues
- mark issues as done
- create issues
- move issues to a different state
- list workflow states (I don't even know what this is)

Tasks are shareable. Now I can give this to my team or anyone. No extra MCP to install. Maybe your team's linear workflow is slightly different. That's fine, just have Nash update the task to meet your needs.

With the rapid progress of these models, I just don't buy the idea we'll have thousands of one-off MCPs to interact with services.

https://www.loom.com/share/c7d6c29a433e4f57b6bfcde94b9200e3?sid=5ea5e324-c9eb-40e9-925b-e708f3370552