The catalogue can store roles and discovery domains while sharing existing skill identities across domains. Initial domain labels can be seeded repeatedly without creating duplicates, and additional domains remain supported.
Posting support is stored once per role–skill relationship and distinct posting. The evidence reader counts successfully extracted postings, applies the same seniority filter to both counts, and returns a bounded set of supporting references. These counts describe the collected sample, not the whole job market.
This adds storage and evidence retrieval. Scheduled posting collection, extraction and personalised path generation remain under development.