PRISM now closes the loop end-to-end with no manual steps: the live collector pulls real dark-web data, the refinement pipeline turns it into a single typed, provenanced ontology (actors, wallets, .onion sites, victims, CVEs — every node tagged with who told us and when), and the products layer publishes a confidence-gated intelligence feed (JSON/CSV), a data-sheet manifest, and per-customer exposure reports. It all runs hands-off inside prism watch — collect → refine → score → publish — built and verified on live data with 187/187 tests green and doctor OPERATIONAL across every layer.