Live Database Ingest

Turns raw database schema snapshots into semantic layer sources and wiki documentation. You point it at JSON files containing table definitions, foreign keys, and connection metadata, and it writes validated sources with proper joins, preserves database comments as descriptions, and creates wiki pages only when there's real business meaning in the schema. The identifier verification protocol is the interesting bit: it actually probes the warehouse with discover_data and entity_details before writing anything, so you don't end up with semantic layer sources that reference tables or columns that don't exist. It won't invent measures or business logic that isn't already in your schema comments, which keeps the output grounded in what's actually there.

GitHub stars: 1.1K
last updated: Jun 11, 2026

View on GitHub

Install

$ npx -y skills add kaelio/ktx --skill live_database_ingest --agent claude-code

Installs into .claude/skills of the current project.

Files

Live Database Ingest

Use this skill when the ingest work unit contains raw files under raw-sources/<connectionId>/live-database/<syncId>/.

Workflow

Read the table JSON file listed in the work unit.
Read connection.json to understand the snapshot metadata.
Read foreign-keys.json when the table has a foreign key or when joins are needed for the semantic-layer source.
Create or update one semantic-layer source for the table with sl_write_source.
Use the physical table name from the raw JSON as the source table field.
Preserve database comments as descriptions.db on tables and columns.
Add joins only when the foreign key index names both sides.
Write wiki pages only for durable business meaning that is present in table or column comments.
Run sl_validate for the table source before the work unit completes.