Current Claude Code level: 2-3 agents on independent tasks. Any more and I’m the bottleneck.
My day is usually: think of 3 things I want to get done, plan each with an agent in herdr, coffee while they research, solidify plans, caffeinate + /rc, go to lunch, babysit on mobile, verify & merge in the afternoon.
Curious what flows are working well for others. Sadly, I’m not a loop writer yet
This week, I automated our bug triage process with Claude Managed Agents. It auto-fixed two bugs first day on the job 🤩
Slack/Anthropic Webhooks + Hono on CloudRun + Managed Agent w/ memory
A new claude code feature that you may have missed is "/tui fullscreen" Enable this! It runs much better and adds mouse support plus this great "Jump to bottom" button 🤩
Great talk from @trq212 today. Favorite idea: ask claude to interview you before you build. AskUserQuestion, in-depth, hunting for ambiguity. You end up with a solid spec easier than trying to write the full plan prompt yourself.
Excited to be a part of anthropic's first dev conference! been building ROMie with claude for months and really excited to see what's next. #CodeWithClaude
Things I've basically stopped doing because of Claude skills:
* writing commit messages
* writing GitHub issues
* triaging bug reports
Curious how this list will grow in the next six months