Same here! That's why I wrote this up – I've found that there are ways to run parallel agents that don't result in the terrifying mess that I thought they would (mainly for research tasks that won't commit code)
Running Parallel Agents for Research Tasks Without Chaos
By
–