I’ve been using Claude Code for 6+ months and started using skills to standardize certain things. But multi-agent teams? That’s where I stopped. Feels like complexity increases dramatically and I still want to be able to follow what’s happening in the code. Is it a losing battle?
There’s so much hype around AI agent teams working together like a dev team. But I’m wondering how it actually works in practice, not just in the success stories. How do you maintain oversight when multiple agents are generating code in parallel?
I know I need to test it myself to really have an opinion. Anyone here who’s tried it and can share experiences – both good and bad?
Originally posted on LinkedIn.