type: initiative slug: quickstart-dashboards title: Quickstart dashboards workstream: mcp-analyst-agent owner: data-ai-engineer-architect status: in_progress milestone: m2-internal-adoption-design-partners tasks: - quickstart-dashboards-program-setup - quickstart-dashboard-pack-salesforce-pilot - quickstart-dashboard-pack-zendesk-pilot
{{ initiative_progress_bar("mcp-analyst-agent", "quickstart-dashboards") }}
Build and maintain opinionated Dataface dashboard packs for Fivetran quickstart open-source dbt projects. Packs should feel familiar to anyone who has used the source product’s standard analytics: same kinds of questions, similar topologies (overview vs entity detail), and sensible cross-board navigation—all backed by the quickstart star schema (no pretend metrics the model cannot support).
Intentionally M2 / design-partner work: it assumes dft init, catalog context, and analyst workflows exist; the goal is repeatable content + navigation patterns on real quickstart repos, not one-off YAML.
Link-heavy packs need first-class board linking before the program is fully unblocked:
Design research: Dashboard linking — research.
Use a dedicated parent directory for clone-per-connector checkouts (recommend outside the Dataface git tree to avoid nested-repo friction). Document the exact path in Spec. Each connector gets a sibling checkout; run dft init at the project root per repo conventions, then author dashboards alongside that project.
.codex/skills/quickstart-product-dashboard-research/SKILL.md — vendor defaults, question inventory, entity/detail boards, link matrix, model gaps.dft init checklist (program setup task).tasks (spawn follow-up tasks as needed).