New here? Pick a path in the project overview: govern and promote agents starts with the Dev-to-Prod Walkthrough; govern tables only starts with Govern tables only.
- Dev-to-Prod Walkthrough — the champion flow: one agent, end to end.
- Import a Genie Agent from UI into Code — import a UI-built Genie agent, then govern it via the walkthrough.
- Quickstart — create a Genie agent from scratch.
- Agents — data citizen: add, update, promote and remove an agent.
- Governance — governance engineer: masks, classes, tiers,
table_readers, row filters,ungovern, acknowledgements. - One team doing both — both kinds of change in one PR.
- Prerequisites — OS, Python, Terraform, network, Databricks account, cloud credentials.
- Pipelines — what to commit, CODEOWNERS,
bootstrap-pipelines, OIDC, the GitHub Actions and Azure DevOps samples. - Deployment machine — releasing prod in v1: persistent checkout,
GENIERAILS_STATE_DIR, locks,release-planandrelease. - Architecture — layers, artifact ownership, config files, Genie agent lifecycle.
- Access —
table_readers,acl_groups,group_map, the deployer SP. - Scheduled governance — scheduling maintain and the audits yourself; the legacy-only (deprecated) scheduled job.
- Advanced Usage — IdP-synced groups, independent environments, masking-UDF reuse.
- Mask library reference — every treatment and
tag_treatments.
- Effective-Access Verification — prove masking/row filters take effect by querying as per-tier principals.
- Integration Testing — unit tests, integration scenarios, test data.
- Troubleshooting — imports, provider quirks, common errors.
- Legacy mode — the older AI-drafted mode, as a migration source only; migrating. Legacy-only: country and industry overlays.
← Back to the project overview.