A practical guide to running the agents on Snowflake Cortex Code — from installing the tooling, to producing a governed, reviewed, production-ready pipeline. OneData isn't a one-shot prompt; it's a staged delivery system with approval gates at every step.
OneData turns business rules and technical metadata into a governed data pipeline — mappings, Snowflake code, dbt assets, and validation scorecards — running as a team of specialist agents inside Snowflake Cortex Code. The difference from a plain "write me some SQL" prompt is control: every phase produces a reviewable artifact, and nothing advances without a human sign-off.
You provide the inputs. OneData produces the outputs. Drop your business requirements and metadata into the project's input/ folder; the agents create and update everything under output/ — with approval gates and validation along the way.
Work through them in order the first time, then use them as reference
Everything that must exist before you run: the Snowflake account, roles and authentication, installing Cortex Code CLI, and configuring your connection profile.
Open setupThe phase-by-phase flow — what each agent does, what it asks you to approve, and what it produces. Plus the folder structure and the four engagement scenarios.
Open run guideThe rules layer that makes the output enterprise-ready — security guardrails, responsible-AI alignment, naming standards, and safe Snowflake patterns.
Open governanceEach phase is owned by a specialist agent — click through to see any of them in detail
Start with setup and installation, or talk to the team about a guided walkthrough on your own environment.