Refactor the capstone for production hardening

The capstone pipeline logic works, but production reviewers want reusable transform code, defensive checks when the API misbehaves, structured logging, and canvas names that read like a runbook.

You'll import the capstone starter, move Transform Data into a dedicated sub-workflow, guard the HTTP Request with a Validate API Response If node plus Stop and Error, layer Data Table checkpoints into an execution_log table, and rename every node for clarity.