What's the last thing your agent still hands back to you?
We build agents all day, and the thing I keep noticing is that they almost never fail in the middle. They fail at the very last step, when something has to happen in the real world.
Ours kept dying in the same spot: a signup form that texts a six-digit code. Everything up to that point ran fine, and then it just sat there waiting for a human to go and read a text message.
I'm curious what everyone else's version of this is. Not the thing your agent does badly - the thing it does perfectly, right up until it has to stop and hand it back to you.
Verification codes? A booking that only happens by phone? A supplier who answers email once a week? Something that needs an actual human voice on the other end?
Mostly I want to know whether the wall sits in the same place for everyone, or whether it moves depending on what you're building.
Replies
The last thing mine hands back is a database migration.
Everything git can undo, I handed over. Refactors, tests, whole features, the agent writes and I read the diff. Running a migration against real data is mine every time. So is the commit, it never commits for me, and not because the messages are bad. It's the last point where I'm forced to actually look at what happened