PPTB Skills

Appearance

add-dataverse-api ​

This skill automates adding window.dataverseAPI calls to a PPTB tool — the HTTP client for CRUD operations, FetchXML/OData queries, entity/attribute/relationship/option-set metadata, custom action and function execution, and solution deployment against Dataverse. A developer reaches for it whenever the tool needs to read, write, or reshape Dataverse data or schema.

Every metadata-modifying call this skill generates — creating or updating an entity, attribute, relationship, or option set — is followed by a call to dataverseAPI.publishCustomizations(), since those changes don't take effect in the environment until published.

When to use it ​

  • "Create/retrieve/update/delete a record from Dataverse in my tool"
  • "Run a FetchXML query" / "run an OData query with filters"
  • "Create a custom table with these columns"
  • "Add a choice field / lookup field / relationship to this table"
  • "Call this custom action" / "call this custom function"
  • "Deploy this solution zip from my tool"
  • "Get metadata for this entity" / "list all entities"

Inputs ​

  • Which operation category is needed: CRUD, relationship association, query (FetchXML/OData), metadata read, entity/attribute/relationship/option-set write, action/function execution, or solution deployment
  • Entity logical name(s) and the columns/attributes involved
  • For writes: the record shape, or the full attribute/entity/relationship/option-set definition (schema name, display name, type, required level, etc.)
  • Whether the call should target the primary or secondary connection (connectionTarget)
  • For custom actions/functions: operation name, operationType ('action' | 'function'), bound entity name/id if bound, and parameters

What it does ​

  1. Adds CRUD calls — dataverseAPI.create(entityLogicalName, record, connectionTarget?), .retrieve(entityLogicalName, id, columns?, connectionTarget?), .update(...), .delete(...), .createMultiple(entityLogicalName, records, connectionTarget?), .updateMultiple(...) — requesting only the columns actually needed on retrieve(), and including @odata.type on every record passed to createMultiple/updateMultiple (plus id for updates).
  2. Adds relationship association calls — dataverseAPI.associate(...) / .disassociate(...) — for many-to-many relationships.
  3. Adds query calls — dataverseAPI.fetchXmlQuery(fetchXml, connectionTarget?) for FetchXML, dataverseAPI.queryData(odataQuery, connectionTarget?) for OData ($select, $filter, $orderby, $top, $skip, $expand).
  4. Adds metadata reads — dataverseAPI.getEntityMetadata(...), .getEntityRelatedMetadata(entityLogicalName, relatedPath, ...), .getAllEntitiesMetadata(...), .getEntitySetName(logicalName), and .getCSDLDocument(connectionTarget?) for the full CSDL/EDMX document when the tool needs to enumerate custom actions/functions or inspect entity shapes.
  5. Adds entity/attribute/relationship/option-set writes — createEntityDefinition, updateEntityDefinition, deleteEntityDefinition, createAttribute, updateAttribute, deleteAttribute, createPolymorphicLookupAttribute, createRelationship, updateRelationship, deleteRelationship, createGlobalOptionSet, updateGlobalOptionSet, deleteGlobalOptionSet, insertOptionValue, updateOptionValue, deleteOptionValue, orderOption — using dataverseAPI.buildLabel(text, languageCode?) for every Label/DisplayName field and dataverseAPI.getAttributeODataType(attributeType) for every attribute's @odata.type.
  6. Appends await dataverseAPI.publishCustomizations(tableLogicalName?, connectionTarget?) immediately after every metadata-modifying call in step 5 (entity deletion is the one exception — it takes effect immediately and needs no publish call).
  7. Adds dataverseAPI.execute(request, connectionTarget?) for custom actions/functions, setting entityName/entityId for bound operations and omitting them for unbound ones.
  8. For solution deployment, adds dataverseAPI.deploySolution(solutionContent, options?, connectionTarget?) followed by a polling loop on dataverseAPI.getImportJobStatus(importJobId, connectionTarget?) rather than assuming the import succeeded.

Example ​

Before: no schema for a "Project" table.

After:

typescript
const newEntity = await dataverseAPI.createEntityDefinition({
  '@odata.type': 'Microsoft.Dynamics.CRM.EntityMetadata',
  LogicalName: 'new_project',
  DisplayName: dataverseAPI.buildLabel('Project'),
  DisplayCollectionName: dataverseAPI.buildLabel('Projects'),
  OwnershipType: 'UserOwned',
  Attributes: [
    {
      '@odata.type': 'Microsoft.Dynamics.CRM.StringAttributeMetadata',
      SchemaName: 'new_Name',
      DisplayName: dataverseAPI.buildLabel('Project Name'),
      RequiredLevel: { Value: 'ApplicationRequired' },
      MaxLength: 100,
    },
  ],
})

await dataverseAPI.publishCustomizations('new_project')

Checklist it enforces ​

  • [ ] retrieve() and getEntityRelatedMetadata() calls request only the columns/properties actually needed.
  • [ ] createMultiple() / updateMultiple() records include @odata.type (and id for updates).
  • [ ] Every metadata-modifying call (createEntityDefinition, createAttribute, createRelationship, createGlobalOptionSet, option value operations, and their update/delete counterparts) is followed by dataverseAPI.publishCustomizations().
  • [ ] Multi-connection tools pass 'secondary' explicitly where the secondary environment is intended.
  • [ ] Solution imports poll getImportJobStatus() until the job completes rather than assuming success.
  • Dataverse API — the API reference this skill is built on
  • Related skills: ../add-toolbox-api/, ../add-powerplatform-api/