cognitive_synergy/__init__.py uses explicit imports — deliberate, or vestigial?

[egoran2]

The CLAUDE.md file calls this out specifically: "The cognitive_synergy/__init__.py uses explicit imports — when adding new public symbols, you must add corresponding import lines there." Most other modules I've poked at (consciousness, knowledge_graph, safety) seem to either re-export from a curated __all__ or rely on the package being installed and submodules being imported lazily.

A couple of things I'm curious about:

1. Was this a deliberate decision (e.g., to avoid circular imports between synergy_metrics, cognitive_engine etc.), or is it a vestige of an earlier layout?
2. If deliberate — would it be worth adding a short comment at the top of cognitive_synergy/__init__.py so future contributors don't "fix" it by switching to wildcard re-exports and then break tests?
3. Is there a documented rule of thumb for which modules should use explicit imports vs which shouldn't? It looks like the boundary right now is "modules that have heavy submodule interdependence" but I'm guessing.

Mostly asking because I almost added a re-export the wrong way and got bitten before re-reading CLAUDE.md.