setup-datamodel

Pass

Audited by Gen Agent Trust Hub on Apr 29, 2026

Risk Level: SAFE
Full Analysis
  • User-In-The-Loop Approval: The skill enforces explicit user confirmation of the proposed data model in Phase 4 before any modifications are made to the Dataverse environment, ensuring human oversight for all schema changes.
  • Authentication and Token Management: The skill securely manages Azure CLI and PAC CLI authentication tokens to interact with official Dataverse OData endpoints. This access is scoped to the user's authorized environments and utilizes standard authentication flows.
  • Automated Validation Hook: A dedicated validation script (validate-datamodel.js) verifies the integrity of the operation by querying the environment to confirm that all tables and columns defined in the manifest were correctly created.
  • Structured Input Processing: The skill can ingest user-provided ER diagrams and Mermaid syntax to generate proposals. This process includes a parsing stage where the agent translates visual or textual data into structured Dataverse definitions for user review.
  • Resilient Execution Patterns: The skill uses localized scripts (dataverse-request.js, verify-dataverse-access.js) to handle API interactions, which include built-in logic for token refreshing and error reporting without automated rollbacks, maintaining a transparent state for the user.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 29, 2026, 11:50 PM