Extracts durable wiki knowledge from staged Sigma data model specs and workbook summaries.
npx skills add https://github.com/Kaelio/ktx --skill sigma-ingest
Please help me install this Agent Skill. Skill: sigma_ingest Source: https://github.com/Kaelio/ktx/tree/main/packages/cli/src/skills/sigma_ingest Command: npx skills add https://github.com/Kaelio/ktx --skill sigma-ingest
Turning Sigma BI workspace content into reusable warehouse knowledge is manual and error-prone. This Skill converts staged Sigma data model specs and workbook summaries into durable ktx wiki candidates, so agents can query the warehouse using accurate business definitions.
data-models/<id>.json specs and captures metric definitions, domain conventions, and business semantics from element names and column descriptions, skipping hidden elements.entity_details and sql_execution probes using connection mappings before writing anything, marking unverifiable references as unverified.sigma-data-models and sigma-workbooks WorkUnits to produce wiki candidates for concepts like Net ARR and churn cohorts without touching the already-projected semantic-layer sources.Load the sigma_ingest skill for the sigma-data-models WorkUnit and write wiki candidates for the business concepts found in the staged Sigma files.
High-intent search queries and answers about installing and using this skill.
Load this Skill for WorkUnits with unitKey sigma-data-models or sigma-workbooks. It reads each staged JSON file, extracts business semantics from element names and column descriptions, and writes wiki candidates with context_candidate_write.
No. The deterministic project() step writes all semantic-layer YAML for warehouse-table elements before this Skill runs. The Skill only reads existing SL sources via sl_read_source and references them in sl_refs.
The Skill maps the element's Sigma connectionId through connectionMappings in sigma-projection-config.json, then calls entity_details to confirm the identifier resolves. Unresolvable identifiers are probed with sql_execution or wrapped as unverified with the raw source path.
The staged set is treated as a recent-changes slice, not the full workspace. The Skill does not infer that absent workbooks were deleted and keeps the workbookFilter settings in scope while processing.
Hidden elements and hidden columns are skipped, along with visualization settings, owner names, version numbers, and workbooks with no durable business semantics such as Untitled Workbook. Derived table elements without a warehouse path are wiki-only.
sigmacomputingMigrate Sisense data models and dashboards to Sigma workbooks with data parity.
AradotsoBuild a semantic context layer for AI agents querying data warehouses.
NirajkashyapIngest documents into an Obsidian wiki as interconnected wiki pages.
PacvueTranslate natural language database questions into safe SELECT SQL for MySQL, ClickHouse, and Doris.