dlt-bootstrap

Set a project up for dlt pipeline development the house way: install dltHub's official AI Workbench project-scoped, then layer Relentless Data conventions on top as an always-on rule. The workbench's toolkits own pipeline building — never re-teach or duplicate what they cover. After bootstrap, step aside: day-to-day work runs through the workbench's own entry points (/find-source, /explore-data, ...), with the house rule applying automatically because it is a rule, not skill-mediated.

First: check for the house rule

Search the project for a committed dlt-house-conventions.md.

• Present → incremental mode: read its frontmatter (source_types, toolkits_installed), then do only what's missing — typically installing an additional toolkit for a new source type and updating the frontmatter.
• Absent → full bootstrap (below).