Incremental data-job loading through standard writes

Context

Imports must retain ordinary record safeguards and make partial outcomes inspectable.

Decision

Bulk custom-record and link loading uses the Data Job Engine, preflight impact validation and ordinary persistence writes, retaining committed batches after later failure.

Alternatives Considered

Direct-table import adapters: A separate write path could bypass naming, validation, attachments, audit and events. Automatic job rollback: Outcomes do not retain the full prior values needed for faithful reversal; reset and reload is the documented recovery route.

Consequences

Reloaded matching lines produce unchanged outcomes where their content still agrees. Partial link correction files can clear attributes. The pinned adapters run engine batches in process rather than through a separate queue consumer.