Viewing profile — adkornev

[dead]

[dead]

What do you do when an agent needs a human decision mid-run? Does Cronloop notify you somehow?

What ended up being the hardest part of running multiple coding agents at once — isolation, figuring out what needs attention, or coordinating dependencies between tasks?

This looks neat. Does the “just import a server function” model stay simple once you add auth, caching and transactions, or does the client/server boundary start leaking through?