Setup Datamodel

What the source does

• FilesystemDeclared1 observation
• NetworkNone observedNo outbound calls in the source
• ShellNone observedNo commands run in the source
• DependenciesPinnedNothing floats
• SecretsNone foundNo keys in source

Five audit layers

Gates applied: no behavioural pass.

const { approve, block, runValidation, findPath, getAuthToken, makeRequest, getEnvironmentUrl } = require('../../../scripts/lib/validation-helpers');

Audit 0.4.1 · source sha 8075ef5705c5 · 18 Oct 2026Report an issue or request a re-scan

• Publishermicrosoft GitHub organizationOwner of the repository this skill is read from
• Source foldermicrosoft/power-platform-skills/plugins/power-pages/skills/setup-datamodelThe exact folder we audited
• LicenceMITAs the repository states it
• Commit pinned for this verdict8075ef5705c5A new audit runs when the source changes
• Repository statusNot archivedPushed today
• Sandboxed runNot runSkills are read, not executed; layer L4 is skipped

Its description: Creates Dataverse tables, columns, and relationships for a Power Pages site based on a data model proposal. Use when the user wants to set up the data model, create database tables, or build the Dataverse schema for…

---
name: setup-datamodel
description: >-
  Creates Dataverse tables, columns, and relationships for a Power Pages site based on a data
  model proposal. Use when the user wants to set up the data model, create database tables,
  or build the Dataverse schema for their site.
user-invocable: true
allowed-tools: Read, Write, Bash, Grep, Glob, AskUserQuestion, Task, TaskCreate, TaskUpdate, TaskList, mcp__plugin_power-pages_microsoft-learn__microsoft_docs_search, mcp__plugin_power-pages_microsoft-learn__microsoft_code_sample_search, mcp__plugin_power-pages_microsoft-learn__microsoft_docs_fetch
model: opus
---

> **Plugin check**: Run `node "${PLUGIN_ROOT}/scripts/check-version.js"` — if it outputs a message, show it to the user before proceeding.

# Set Up Dataverse Data Model

Guide the user through creating Dataverse tables, columns, and relationships for their Power Pages site. Follow a systematic approach: verify prerequisites, obtain a data model (via AI analysis or user-provided diagram), review and approve, then create all schema objects via OData API.

## Core Principles

- **Never create without approval**: Always present the full data model proposal and get explicit user confirmation before making any Dataverse changes.
- **Use TaskCreate/TaskUpdate**: Track all progress throughout all phases — create the todo list upfront with all phases before starting any work.
- **Resilient execution**: Refresh tokens proactively, check for existing tables before creating, and report failures without automated rollback.

**Initial request:** $ARGUMENTS

---

## Phase 1: Verify Prerequisites

**Goal**: Confirm PAC CLI authentication, acquire an Azure CLI token, and verify API access

**Actions**:

1. Create todo list with all 8 phases (see [Progress Tracking](#progress-tracking) table)
2. Follow the prerequisite steps in `${PLUGIN_ROOT}/references/dataverse-prerequisites.md` to verify PAC CLI auth, acquire an Azure CLI token, and confirm API access. Note the environment URL as `<envUrl>` for subsequent script calls.

**Output**: Verified PAC CLI auth, valid Azure CLI token, confirmed API access, `<envUrl>` noted

---

## Phase 2: Choose Data Model Source

**Goal**: Determine whether the user will upload an existing ER diagram or let AI analyze the site

**Actions**:

<!-- gate: setup-datamodel:2.source | category=plan | cancel-leaves=nothing -->

> 🚦 **Gate (plan · setup-datamodel:2.source):** Decide whether the user uploads an existing ER diagram or the data-model-architect agent infers the model. Choice routes the rest of the skill into Path A vs Path B.
>
> **Trigger:** Entering Phase 2.
> **Why we ask:** Auto-picking either path can run a multi-minute architect agent against the wrong intent (Path B) or skip Dataverse-existence checks (Path A).
> **Cancel leaves:** Nothing — no Dataverse calls made yet.

1. Ask the user how they want to define the data model using the `AskUserQuestion` tool:

   **Question**: "How would you like to define the data model for your site?"

   | Option | Description |
   |--------|-------------|
   | **Upload an existing ER diagram** | Provide an image (PNG/JPG) or Mermaid diagram of your existing data model |
   | **Let the Data Model Architect figure it out** | The Data Model Architect will analyze your site's source code and propose a data model automatically |

2. Route to the appropriate path:

### Path A: Upload Existing ER Diagram

If the user chooses to upload an existing diagram:

1. Ask the user to provide their ER diagram. Supported formats:
   - **Image file** (PNG, JPG) — Use the `Read` tool to view the image and extract tables, columns, relationships, and cardinalities from it
   - **Mermaid syntax** — The user can paste Mermaid ER diagram text directly in chat
   - **Text description** — A structured list of tables, columns, and relationships

2. Parse the diagram into the same structured format used by the data-model-architect agent:
   - Publisher prefix (ask the user, or retrieve from the environment via `pac env who`)
   - Table definitions: `logicalName`, `displayName`, `status` (new/modified/reused), `columns`, `relationships`
   - Column definitions: `logicalName`, `displayName`, `type`, `required`
   - Relationship definitions: type (1:N or M:N), referenced/referencing tables

3. Query existing Dataverse tables (same as Phase 3 would) to mark each table as `new`, `modified`, or `reused`.

4. Generate a Mermaid ER diagram from the parsed data (if the user provided an image or text) for visual confirmation.

5. Proceed directly to **Phase 4: Review Proposal** with the parsed data model.

### Path B: Let the Data Model Architect Figure It Out

If the user chooses to let the Data Model Architect figure it out, proceed to **Phase 3: Invoke Data Model Architect** (the existing automated flow).

**Output**: Data model source chosen and, for Path A, parsed data model ready for review

---

## Phase 3: Invoke Data Model Architect

**Goal**: Spawn the data-model-architect agent to autonomously analyze the site and propose a data model

**Actions**:

1. Use the `Task` tool to spawn the `data-model-architect` agent. This agent autonomously:
   - Analyzes the site's source code to infer data requirements
   - Queries existing Dataverse tables via OData GET requests
   - Identifies reuse opportunities (reuse, extend, or create new)
   - Proposes a complete data model with an ER diagram

2. Spawn the agent:

   ```
   Task tool:
     subagent_type: general-purpose
     prompt: |
       You are the data-model-architect agent. Follow the instructions in
       the agent definition file at:
       ${PLUGIN_ROOT}/agents/data-model-architect.md

       Analyze the current project and Dataverse environment, then propose
       a complete data model. Return:
       1. Publisher prefix
       2. Table definitions (logicalName, displayName, status, columns, relationships)
       3. Mermaid ER diagram
   ```

3. Wait for the agent to return its structured proposal

Excerpt, first 6,001 characters of SKILL.md, verbatim from the audited commit. It is text from someone else’s repository, quoted so you can judge it before your agent reads it.

git clone https://github.com/microsoft/power-platform-skills
git -C power-platform-skills checkout 8075ef5705c5

Works in

1. 2026-10-18
   ✓SAFE B 89 / 100

   first audit · source 8075ef5705c5

Questions about Setup Datamodel

• github.com/microsoft/power-platform-skills/tree/main/plugins/pow — the audited source at commit 8075ef5705c5 and every finding aboveObserved · 18 Oct 2026

Spotted something wrong? Report a correction · Methodology