question for people running coding agents unattended:
how do you know what they did while you were away? i read the diffs, but the reasoning behind them is gone by morning.
session logs, summaries, a dashboard, nothing? curious what actually works for you 👇
Commit
8b2f3d8
Pretty much similar to mine as I look at the diff, but when it comes leaving my agents unattended, I only execute low risk (chore like) tasks
low risk chores only is the sane line. what counts as low risk for you, anything that can't touch prod?
commit 8b2f3d8
yes, pretty much non-prod stuff. You can say that i still doubt the abilities of ai coding agents :)