Split non-database responsibilities out of Model

[davorg]

Problem

Succession::Model currently combines several different responsibilities:

• DBIx::Class queries and page view data;
• genealogy graph management;
• remote blog-feed fetching and fallback caching;
• shop JSON loading and HTTP validators;
• reference Markdown menu discovery;
• diagnostics and curated historical data.

The file has now been organised and documented, but its size and dependency surface will continue growing if every data-related feature lands there.

Proposed work

Extract responsibilities incrementally rather than in a large rewrite. Good first candidates are the isolated, non-database services:

1. blog feed fetching/caching;
2. shop JSON loading;
3. reference-menu discovery.

Keep route-facing behaviour stable and inject the extracted services so their existing tests remain deterministic.

Acceptance criteria

• Agree a small service boundary and naming convention before extraction.
• Extract one responsibility at a time with no route behaviour change.
• Succession::Model trends toward database/domain operations rather than generic application I/O.
• Existing feed, shop, query-budget, and full-suite tests continue to pass after each extraction.