← sources

Loops/Next-Task -> ?? || What’s next?

source: reddit (reddit:ClaudeCode)published 2026-11-01 01:10ingested 2026-11-02 02:41

retained text

I’ve been building a b2c SaaS for the live music industry for a few months with Claude code (along with the guidance of a senior dev who happens to be my best friend/cto).

I’ve developed an agent team of product/pm/front-end/back-end/devops/design. They’ve taken a lot of guidance from me and it’s building exactly what I want.

We’re getting close to launching in beta, and I’ve been using loop/next-task and it’s been a great way to get through 550 PRs which work when pushed to our dev environment in AWS. 

While I can leave it running on its own machine, and it can work for days uninterrupted it’s very much a single track team (pm selects task, then prod sets requirements, then front end, then DevOps pushes to GitHub) it seems like it’s very one track. Like, I can’t get it to scale where some times it’s a single back end agent taking two hours to complete a task, and I feel like it could be running swarms of multiple on each step to speed things up.

Is this a setting/prompting problem, or something that’s in my agent design I could mess with?

Or is there a better way to get the team working than loop/next-task?

document fetched 2026-11-02 00:36 · canonical https://reddit.com/r/ClaudeCode/comments/1xkqgqc/loopsnexttask_whats_next