dataverse-eppcdemo2fre-mcp

name	dataverse-eppcdemo2fre-mcp
description	Query and update Microsoft Dataverse records in the EppcDemo2FRE environment (https://eppcdemo2fre.crm.dynamics.com/api/mcp_preview) via the preview Dataverse MCP endpoint. Read, search, create, and update rows in standard Dataverse tables (account, contact, lead, opportunity, incident) and custom tables in this environment. Trigger phrases: - "look up the [record] for [name]" - "show me open [records]" - "find the [record] [name] in Dataverse" - "create a new [record] in Dataverse" - "update the [field] of [record]" - "what [records] are assigned to me" - "search Dataverse for [keyword]" - "list rows in the [table] table" Scope: the EppcDemo2FRE environment ONLY. Do NOT use for other Dataverse environments, SharePoint lists, Excel files, or Power BI artifacts.
license	MIT
metadata	{"author":"Launch Control","version":"1.0"}

Dataverse MCP: EppcDemo2FRE

Connects Microsoft 365 Copilot Cowork to the EppcDemo2FRE Dataverse environment through the native Dataverse MCP server (connector dataverse-eppcdemo2fre). Authentication flows through Cowork's OAuthPluginVault: users sign in once with their own Dataverse identity and Dataverse security roles enforce access. This plugin instance is scoped to EppcDemo2FRE only; for other environments, install a separate plugin instance.

When NOT to Use

• Other Dataverse environments: install a separate env-scoped plugin.
• Logical-name discovery or solution dumps: use dataverse-eppcdemo2fre-schema.
• Business policy questions (valid status transitions, approvals, escalation): use dataverse-eppcdemo2fre-business-skills, which loads the authoritative policy rows from the Dataverse skill table.
• SharePoint lists, Excel / CSV, or Power BI: use the matching native tool.

Core Concepts

• Table: the Dataverse equivalent of a SQL table. Standard tables include account, contact, opportunity, lead, incident (case). Custom tables are named per the environment.
• Row: a single record, identified by a GUID primary key.
• Column: a single property on a row.
• Lookup: a relationship column pointing to a row in another table.
• Choice: an enum-style column with a fixed set of integer-backed values.

Tool Use: preview MCP (search / describe / execute)

You have access to the Dataverse MCP server backed by the preview endpoint https://eppcdemo2fre.crm.dynamics.com/api/mcp_preview. Every operation is routed through one of three tools via a filesystem-style path:

• search: keyword search across the environment. Returns paths for table schemas (tables/<name>), Business Skills (skills/<name>), and Custom APIs (api/<name>). Use first to discover; never hard-code paths.
• describe: full details for any path returned by search.
  • describe("tables/"): list every table in the env.
  • describe("tables/<name>"): full schema (columns, types, relationships, example queries). Call before querying when column names are unknown.
  • describe("tables/<name>/records/<guid>"): full single record.
  • describe("skills/<name>"): full Business Skill body.
  • describe("api/<name>"): Custom API input parameters and response shape.
• execute: perform an operation against a path. The verb is one of:
  • read: SQL SELECT against tables/. Supports SELECT, TOP, WHERE, ORDER BY, GROUP BY, JOIN, aggregates. Does not support subqueries, HAVING, DISTINCT, UNION, OFFSET. Always pass a TOP bound (default 25 to 100, max 1000).
  • create / update / delete: row CRUD on tables/<name>/records[/{guid}]. Destructive operations require hasUserApproved: true in the query body.
  • Custom API invocation: see below.

Invoking Custom APIs (preview MCP)

Some workflows depend on Dataverse Custom APIs (unbound actions). The preview MCP exposes them through the same three tools:

1. search("<intent>") surfaces the Custom API path (for example api/<publisher>_<Name>).
2. describe("api/<name>") returns the input parameters (case-sensitive) and response shape.
3. execute runs the action with those parameters.

Rules: never invent an action; only call ones surfaced by search or named in a Business Skill body or user prompt. Always describe before execute. Treat actions as side-effecting. Do not hand-compute a result an action is meant to produce; if it errors, surface the error verbatim.

Workflow

1. Resolve the table the user is asking about. If ambiguous, ask or call search to disambiguate.
2. Resolve people or entity references before querying.
3. describe before querying when columns are unknown. For pure schema questions defer to dataverse-eppcdemo2fre-schema; for policy questions defer to dataverse-eppcdemo2fre-business-skills.
4. execute with operation read using the smallest filter that satisfies the request; prefer server-side WHERE and TOP over fetching everything.
5. Page through when the user asks for "all" rows.
6. For writes, confirm the target row and the changes before executing.

Output Format

• List queries: a compact table with the 4 to 6 most relevant columns.
• Single-row lookups: a key/value summary of important fields plus the id.
• Writes: confirm the operation and surface the new or updated row id and URL.

Confirmation Gates

• Before any create, update, or delete: state the table, the row (by name and id), and the changes. Proceed only after explicit confirmation in the current turn.
• For bulk operations (more than 5 rows affected): always confirm.

Errors

Error	Likely cause	Fix
401 / unauthorized	OAuth token expired or scope missing	User reconnects to the plugin in Cowork
403 / forbidden	Security role lacks privilege on the table	Admin grants the role; not a plugin issue
404 / row not found	Stale id or row deleted	Re-search by name
400 / invalid column	Logical vs display name mismatch	Use the logical name from describe("tables/<name>") or dataverse-eppcdemo2fre-schema
@odata.bind not supported on create / update	Tried OData lookup bind syntax	The preview MCP does not accept @odata.bind; set the lookup attribute directly to the GUID string