question on architecture decision documentation (decisions:19 in repo health scan)

[bmccall17]

hi, we've been researching repo health patterns across the open source ecosystem, building a tool called repOptics to understand how different projects handle things like design documentation. ran a scan on langflow and the decisions category scored 19/100, the lowest in the B-tier of our dataset. the gap is specifically around architecture decision records and documented design rationale. noticed there are 60+ architecture-related issues in the last couple months, which suggests contributors are asking "why" questions regularly. curious whether formalizing some of the core design choices (graph execution model, component plugin architecture) as ADRs has come up before.