Get the threat model for one application

Fetch the customer-editable threat model for one application: the application overview (PersistentSastScanApplicationInfo.llmContext), app-info metadata, the fixed sections (components, interfaces, actor types, authn/authz model, control plane, attacker types, application-specific attacker objectives, datastores, assumptions, out of scope), and any user-added custom sections. Returns threatModel: null for the threat model if the application's first scan hasn't generated one yet — applicationOverview is always returned because llmContext lives on the application row independent of threat-model existence. The application's overview and fixed sections drive ZeroPath's investigation, validation, and triage agents — editing them changes what the next scan finds and how it's triaged.