Give a personal workflow its own service boundary

Keep the entry point focused

The hosted factory mounts the job-applications module at a known path and supplies a dedicated runtime data directory plus the shared credential context. The source collection separates postings, decisions, service handling, storage, and views.

That organization gives the workflow a smaller unit of change. Application-tracking logic can be reviewed and released without pulling unrelated creative or infrastructure tools into the same process.

Personal records are runtime data

The category package excludes changing records, desktop credentials, and private configuration. Hosted data uses the controller's service-scoped storage API. Updates publish a revision-checked manifest, so conflicting edits are surfaced instead of silently lost.

The process binds to loopback and relies on verified owner identity from the controller. Its source manifest records exported modules for review, while a hosted-status surface reports missing dependencies.

A dedicated runtime directory isolates this workflow's records from the other categories. The constructor makes that scope visible at the point where the handler is created.

Keep the workflow bounded

The inspected factory establishes a dedicated application workspace, not automated submission to employer systems. Postings, decisions, storage, and presentation remain separate modules behind one service boundary. That keeps private records in runtime storage and makes the workflow independently reviewable without claiming authority over external accounts or submission outcomes.