ktx Data Agent Context Layer

Skill by ara.so — AI Agent Skills collection.

ktx is a self-improving context layer that teaches AI agents how to query data warehouses accurately. It automatically builds semantic layers from your warehouse metadata, dbt models, BI tools, and company wikis, then exposes that context through MCP (Model Context Protocol) and CLI tools. This enables agents to fetch approved metrics and write better SQL instead of reinventing logic on every query.

What ktx Does

• Builds warehouse context automatically — samples tables, captures metadata, detects joinable columns
• Ingests company knowledge — from dbt, LookML, Looker, Metabase, Notion, and wiki content
• Creates a semantic layer — with join graphs that resolve fan/chasm traps
• Flags contradictions — across different knowledge sources
• Serves agents via MCP — with full-text and semantic search across wiki and metrics
• Read-only by design — never writes to your database

Supports PostgreSQL, Snowflake, BigQuery, ClickHouse, MySQL, SQL Server, SQLite.

Installation