Demo brain. Simulated stack for a fictional client — see onboarding-status.

Stack

One-paragraph purpose: what Umbrella Corporation is built with, and which src/ submodule each part lives in — check this before opening any submodule. See architecture for how these pieces fit together during the migration.

Submodules

None yet. This is a demo brain with no real client repository to add as a src/<RepoName> submodule — see src/README.md. In a real engagement at this delivery stage, src/hive-os (the new platform) would be added once the client grants repo access; src/umbra-legacy would follow if the strangler-fig work needs to read the legacy schema directly.

Languages & frameworks

• Hive OS (new platform) — TypeScript throughout. React frontend; NestJS backend service (facility-access, sample-inventory as separate modules within one service for Phase 1, per decisions); PostgreSQL for all new data.
• UMBRA (legacy, being replaced) — VB.NET WinForms client against an on-prem Oracle 11g database, no API layer; integration during the strangler-fig period is a nightly batch export Marcus Webb’s team runs from Oracle into a staging bucket Hive OS reads from. (source: Granola notes, “Discovery: Facility Access Workflows” 2026-07-02 and “Sprint 1 Planning” 2026-08-06)

Infrastructure

• Hosting: AWS (Umbrella’s existing enterprise agreement), ECS Fargate for the NestJS service, RDS for PostgreSQL.
• Auth: Okta SSO, badge-system integration handled separately by Daniel Osei’s team on Umbrella’s side (out of scope for Hive OS itself, Hive OS only consumes the badge events).
• CI/CD: GitHub Actions, per Loomery’s default pipeline template — not yet confirmed against Umbrella’s own change-control process, which Priya Anand’s team will need to sign off before Phase 1 goes live at the Raccoon City pilot site. (source: Granola note “Sprint 1 Planning”, 2026-08-06)