See Subagents Work Side By Side
From the course: Agentic AI For Founders & Senior Operators
In our weekly stock analyst, three general-purpose subagents ran in parallel, each within a 15-credit budget, and produced a report that's a good start. In Claude Code you can see each subagent, the main agent waiting on them, how many tools each used and the full transcript. Some views hide the transcript, but you can still see which tools each subagent called.
Transcript
So it used three separate general purpose subagents that ran in parallel, each scoped over the 15-credit budget: six, four, seven. All right. All good. And then it also used the built-in browser. So this is a pretty useful report. Obviously, lots to do here in terms of the output, and it would need to go through several iterations of refining it, maybe fixing the inputs, but it's a good start. And you know, you've effectively got a weekly stock analyst that's doing work for you. Let's see what our friend on the code side has come up with. We had given it the same task and it also has started some subagents, which we can actually see here. And you can see that it's got these three agents and this is the main agent that's waiting. And then these two and the number of tools that they've used. You can see the full transcript. So all of that is visible. And that's also in fact visible on the desktop app while those agents are running. Let's see if we're able to find... Okay. In this case, we should be able to see what each agent did. Yep. You don't get to see the transcript. They abstract that away, but you can see at least which tools that they called.