When did your AI coding agent last make a scope or architecture decision you had to undo? I’m building Archseed to help make those decisions explicit before coding. Does that add value, or does your current planning workflow already cover it?