source-survey

What problem does it solve?

Manually surveying source database structures, reconciling design documentation with live schemas, and generating consistent design artifacts is time-consuming, error-prone, and requires deep knowledge of multiple tooling and formatting standards. This skill eliminates that manual overhead for data modelers working on Oracle-based data lakehouse projects.

Core Features & Use Cases

• Full Source Survey: Automatically extract table lists, column definitions, comments, and foreign key constraints from Oracle source databases via MCP to generate standardized UTF-8 CSV files for tables and columns.
• Reconcile Mode: Compare existing design CSV files against live database schemas without overwriting original documentation, flagging table mismatches, unconfirmed foreign keys, and schema conflicts for resolution before LLD work begins.
• Interactive ER Diagram Generation: Create filterable, zoomable HTML ER diagrams grouped by business function, with BRD scope badges, foreign key line highlighting, and drag-and-drop layout persistence for team collaboration.
• BRD Source YAML Generation: Auto-generate compliant BRD Source YAML files with inferred functional groups, ingestion modes, and related table mappings, plus per-table column YAML files aligned with project schemas.
• Use Case: For a data modeler onboarding a new Oracle source system for the UBCK lakehouse project, use this skill to run a full survey to produce initial CSV documentation, ER diagrams, and BRD YAML files in a single workflow.

Quick Start

Use the source-survey skill to run a full survey of the NHNCK Oracle source system, export standardized table and column CSV files, generate an interactive ER diagram, and produce the corresponding BRD Source YAML artifacts.