beans archive should be a real archive

[oliverandrich]

I just used beans archive without realizing that it is not an archive but a clean-up command. It would be cool, to have a real archive for bigger projects. beans list could list the open issues and beans list archive could list the stuff in the archive.

[hmans]

main already contains some work in that direction:

• The code that loads beans from .beans/* will now traverse all subdirectories (this is in preparation for some features around automatically restructuring your beans into a directory structure that reflects milestones/epics)
• beans archive will merely move completed and scrapped beans into .beans/archive/

I've been experimenting with some ideas on top of that (including not loading .beans/archive/ unless you explicitly pass an --include-archive CLI arg), but many of these very quickly went against my goal to keep Beans as simple as possible (where possible).

I'll keep experimenting, so keep that feedback coming, it's very valuable.

[hmans]

Allow me to move this over to discussions, since it's more feedback than an immediately actionable issue. Thanks!

[skazazes-work]

I agree that having a proper 'archive' mechanism would be a great improvement to beans. I've found claude code's general behavior improves when it has historical beans it can read and reference.

Alternatively, toggle allowing filtering completed Milestones/Epics and their children in the TUI would leave the full historical context available, but remove a lot of the noise from the TUI