Manage Your Sub-Agents Effortlessly 🤖

So we're really focusing on trying to make sub agents extremely useful and very easy to work with.

So I want to show you how one MCP called station can manage all of your sub agents and MCP servers.

So a big problem with MCP is that you don't want to completely pollute your working session.

And those are the ones that get inherited by the sub agents.

And what we've really done is separate both.

So that way you can actually create sub agents that get attached to MCP servers without ever polluting your thing.

Now, the other thing is you can actually organize them in environments.

So if you look here, you can see in the default environment, I have a few agents set up.

The Playwright agent, the hello World agent, and then a playwright MCP server.

Now, if you look at my cloud code, all I have connected is actually the Station, MCP and, deep wiki, right?

That way I have agents that are not tied to my cloud directly.

I can actually use them in cursor.

I can.

I, can group them together in different environments.

So I have a security one with different agents and different MCP servers.

And none of that is polluting my actual cloud context.

This makes it very easy to bundle it, to make them into docker containers to share them.

So I can bundle them, make this entirely reproducible.

So if I share with the public, endpoint, I can go ahead and install that bundle.

I got a new security bundle.

Now, if I look in an environment, I can actually unpack those agents and those MCP servers.

So my teammates have the exact same configuration that I have.

this is how we're looking at, MCPs and sub agents and making them 20 times more powerful.