A research-ethics scope lock a build can check. 61 lines of CI enforcement that fail the build when an agent drifts outside approved scope — plus the ADRs, threat model and data classification behind it.
nextjs threat-model architecture-decision-records human-in-the-loop privacy-by-design ai-governance research-ethics agent-governance digital-public-good scope-lock
-
Updated
Sep 7, 2026 - TypeScript