Problem

All flows live in a flat .one/flows/ directory. At scale (50+ flows), this becomes hard to navigate. Flow authors resort to filename-prefix conventions (deal-log, deal-update, deal-status) to group related flows, but this is a convention the tool doesn't understand.

Sub-flow references also use flat keys:

{ "type": "subflow", "subflow": { "key": "company-research" } }

There's no way to express hierarchy or scope.

Proposal

Support subdirectories:

.one/flows/
  research/
    company-research.flow.json
    competitor-research.flow.json
  deal-ops/
    deal-log.flow.json
    deal-update.flow.json
  infra/
    constants-load.flow.json
    cache-clear.flow.json

Sub-flow references could use either the flat key (backward compatible) or a namespaced key:

{ "type": "subflow", "subflow": { "key": "research/company-research" } }

one flow list could group by directory. one flow execute could support both one flow execute company-research and one flow execute research/company-research.

Why this matters

Flow-heavy projects inevitably develop categories. Supporting directories is low-cost (glob **/*.flow.json instead of *.flow.json) and avoids the need for naming conventions that the tool can't enforce or leverage.