Just as an introduction, the agents I use (Claude, Gemini, chatGPT) seem to know Meteor very well! 
I see reports of people using dozens or hundreds of agents. When I run agents in Anti-Gravity, I know it is using 2 or more agents (because sometimes you see them talking to each other), but I don’t think it’s dozens or hundreds. And usually it does a great job.
But often I don’t like the plan it asks for approval of. Or it makes mistakes. And that’s just with 2 or 3 agents.
How can people possibly get good results with hundreds?