eddytor-getting-started

name	eddytor-getting-started
description	Onboards new users to the Eddytor MDM platform via MCP connection, authentication, and first table creation. Activates for Eddytor onboarding, MCP setup, workspace exploration, or first table creation — even without the phrase "getting started."
license	CC-BY-NC-4.0
metadata	{"author":"eddytor","version":"1.0"}

Getting Started with Eddytor

Eddytor is an MDM platform. Interact with it through MCP tools: create tables, import data, enforce constraints, manage master data lifecycle.

Connecting

MCP endpoint: http[s]://[<your_domain>|localhost]/mcp Auth: OAuth 2.1 (Supabase-backed), handled automatically by MCP clients.

For Cursor, add to .cursor/mcp.json:

{ "mcpServers": { "eddytor": { "url": "http[s]://[<your_domain>|localhost]/mcp" } } }

First interaction

1. Call list_tables — see all tables with fully-qualified names, storage paths, last modified
2. Pick a table → get_table_schema for columns/types → get_table_metadata for constraints/PKs/version
3. Read data with query_rows (default) or execute_sql (joins/CTEs only)

Gotchas

• Every tool that takes a table parameter requires the fully-qualified name (catalog.schema.table). Get this from list_tables — never construct it manually.
• table_name in create_table and import_csv is the bare name only (e.g., products), never prefixed.
• location is a full storage URL (e.g., abfss://[e-mail]/path).
• Table names in execute_sql must be backtick-quoted per component: `catalog`.`schema`.`table`. Unquoted dots cause parse errors.
• Every table needs at least one primary key column.

Common first-session workflows

Import existing data:

1. infer_schema with CSV content → review types and domain candidates
2. import_csv → set domains on flagged columns → validate_domain_values

Create from scratch:

1. create_table with column defs (at least one PK)
2. set_column_domain on categorical columns
3. insert_rows

Explore existing data:

1. list_tables → get_table_schema → profile_table → query_rows with limit: 10

Checklist: first table setup

• list_tables to confirm workspace access
• get_table_schema on at least one table
• Successfully query_rows with a filter
• If creating: create_table with PK + merge_rows + validate_constraints

Guidelines

• Always call list_tables first — never assume table names
• Default to query_rows over execute_sql
• Check get_table_history before destructive operations
• Set domain constraints early, before importing data