What problem does it solve?
Setting up ktx, the open-source context layer for data agents, requires driving an interactive CLI that agents cannot operate, plus correctly wiring databases, LLM backends, embeddings, secrets, and agent integrations without silent failures.
Core Features & Use Cases
- Non-interactive setup: Runs
ktx setup --no-input with hidden CLI flags to configure LLM backends, embeddings, and database connections in one scripted flow.
- Context source ingestion: Adds dbt, Metabase, Looker, LookML, MetricFlow, and Notion sources one at a time, then ingests them into wiki Markdown and semantic-layer YAML.
- Agent integration and verification: Installs agent rules for Claude Code, Codex, Cursor, and others, starts the MCP server, and validates readiness via
ktx status --json.
- Use Case: A user asks their agent to connect a Postgres warehouse and a Metabase instance to ktx; the skill runs setup, ingests both connections, installs agent rules, and reports a ready verdict.
Quick Start
Use the ktx skill to install and configure ktx in this project, connect my Postgres database, and ingest its schema.