What problem does it solve?
Enterprise users often cannot find which database table or view holds the data they need, or who governs it. This Skill runs a fixed two-step retrieval pipeline across ontology knowledge networks to answer "where is the table" and "who is responsible for this data" questions with grounded, non-fabricated results.
Core Features & Use Cases
- Metadata KN Retrieval: Searches a metadata-type knowledge network via
bkn object-type query with match + knn conditions to find candidate tables/views, extracting technical_name, business_name, owning department, and subject domain.
- Duty KN Retrieval: Derives department and subject keywords from step 1 results and queries a duty-type knowledge network for governance responsibilities and boundaries.
- Structured Chinese Summary: Merges factual findings (candidate tables with full business names) and governance descriptions (department duties) into a final answer with next-step suggestions.
- Use Case: A data analyst asks "Which wide table holds purchase order data, and which department governs it?" The Skill retrieves candidate tables from the metadata KN, then finds the responsible department's duties, and returns a combined answer.
Quick Start
Ask the assistant to find which table or view contains a specific business data set and which department is responsible for governing it.
