A Next.js + Supabase (TypeScript) platform built around a connector-adapter pattern: every data source — spreadsheets, databases, HCM systems, ERP systems, tax filing, web scraping — implements a shared interface for connecting, extracting, and loading data, registered through a central adapter registry. Visible only to the project team; not published here.

Production build output for the Next.js app, deployed to Vercel, plus any containerized services. Build provenance and deployment logs are tracked internally; images themselves are not published here.

Third-party packages the platform depends on (e.g. cheerio, pdf-parse, exceljs for the web-scraper connector), vetted for known vulnerabilities before adoption. The full dependency manifest is internal.

Feature specs and connector design docs — e.g. the decision to use exceljs over xlsx for the web-scraper connector due to unpatched CVEs in the latter. Internal planning documents, shared with clients on request.

Type-checking (tsc), linting, and connector test coverage run before every release, plus pipeline-run diagnostic logs. Internal to the project team; can include client data and is never published publicly.

Working notes and the record of choices made along the way — e.g. terminology decisions, connector prioritization, rollout sequencing. Internal to the project team.