Area | What to look for |
Dataverse Web API / metadata API | Has created or modified tables, columns, and relationships via direct Web API calls (not only the Maker UI) — e.g. EntityDefinitions, RelationshipDefinitions, PublishXml, solution component registration. Knows a lookup column can't be created through a plain attribute POST and must go through the relationship endpoint. |
Power Automate internals | Has authored flows using Workflow Definition Language (WDL) expressions directly, not only the visual designer. Understands webhook trigger subscription registration and connection references, and has debugged a flow using run-history/API tooling rather than only the Maker Test button. |
Dataverse security model | Hands-on with security roles and table-level privilege scoping (Create/Read/Write/Append/Delete), business units, teams, and just-in-time (JIT) user provisioning via identity impersonation. Has designed or evaluated a dedicated service-account/service-principal pattern, as distinct from running automation under a personal interactive connection. |
Cross-environment integration | Has built a flow or service that calls one Dataverse environment's Web API from automation running in another — including OAuth client-credentials auth and secret management (e.g. Azure Key Vault). |
Power Apps code apps / generative pages | React + TypeScript experience, comfortable with a component library (Fluent UI or equivalent), and understands that a platform's custom data-binding layer may have its own conventions that differ from the raw Web API — getting this wrong can silently orphan records with no error. |
PAC CLI | Fluent beyond basic solution export/import — entity/attribute metadata operations, environment variables, connection references, solution component management. |
Reading exported solutions | Can unpack a solution export (managed or unmanaged) and read its raw contents — classic workflow XAML, Power Automate JSON definitions, sitemap/role XML — to determine what an existing system actually does, as opposed to what its documentation claims. |
Empirical troubleshooting | Comfortable concluding the documentation doesn't define this and resolving it by direct, deliberate testing against a live environment, rather than guessing or stopping at the docs. |
