| name | dataproduct-bootstrap |
| description | Bootstrap a brand-new Snowflake dbt data product from scratch in one go — dbt_project.yml, model layout, README, .gitignore, profiles.yml.example for Snowflake, ODPS, output-port ODCS, openlineage.yml, and a GitHub Actions workflow. Trigger when the user asks to start a new data product, scaffold a new dbt project, or "create a Snowflake data product from scratch." |
Bootstrap a new Snowflake dbt data product (demo)
Scaffold a Snowflake-only dbt data product following the Entropy Data conventions. Demo-grade: one batched question, one platform, no audits, no migration logic.
For populating the contract schema and dbt model bodies from a published data product, use the dataproduct-implement skill after this one.
What this skill produces
After running, the directory contains:
.
├── dbt_project.yml
├── .gitignore
├── README.md
├── profiles.yml.example
├── <data-product-id>.odps.yaml
├── openlineage.yml
├── .github/workflows/data-product.yml
├── macros/
│ └── get_custom_schema.sql
└── models/
├── input_ports/_models.yml
├── staging/_models.yml
├── intermediate/_models.yml
└── output_ports/v1/
├── _models.yml
└── <contract-id>.odcs.yaml
How to run this skill
${PLUGIN_ROOT} below refers to the root of this plugin — the directory that contains skills/. On Claude Code it is set automatically as ${CLAUDE_PLUGIN_ROOT}; use that. On any other agent (Codex, Copilot CLI, etc.) it is unset; resolve it as ../.. relative to this SKILL.md file's directory (i.e. the grandparent of skills/<this-skill>/).
Plan announcement (before Step 1)
Before running Step 1, print this plan to the user verbatim:
Running dataproduct-bootstrap. I'll:
- Pre-check: confirm the working directory is empty.
- Ask you for parameters in one batched question (id, name, purpose, team, Snowflake database, output port table).
- Scaffold the dbt project (dbt_project.yml, profiles.yml.example, model layout, README, .gitignore).
- Scaffold the publishing layer (ODPS, output-port ODCS, openlineage.yml, GitHub Actions workflow).
- Summarize what was scaffolded and the next manual steps.
Then proceed.
Step 1 — Pre-checks
- Confirm the working directory is empty, or contains only files the user is fine with (e.g. an empty git repo, a
LICENSE, or a placeholder README.md that will be overwritten).
- If
dbt_project.yml already exists, stop and tell the user this demo skill only handles greenfield directories. Point them at dataproduct-implement to add models to an existing project.
Step 2 — Gather parameters in one batched question
Ask the user for these in a single prompt. Do not generate any files until you have all of them.
| Parameter | Description | Example |
|---|
DATA_PRODUCT_ID | Stable id, snake_case, also the dbt project name | dp_acme_customer_activity |
DATA_PRODUCT_NAME | Human-friendly name | Customer Activity |
PURPOSE | One sentence — why this data product exists | Customer activity for customer success. |
TEAM_NAME | Owning team id (free-text accepted for the demo) | customer-success |
DATABASE | Snowflake database | ENTROPY_DATA_PROD |
TABLE | First output port table name | customer_activity |
Derive (matching the dataproduct-builder-dbt plugin's conventions, per the dataproduct-dbt skill):
DBT_PROJECT_NAME = DATA_PRODUCT_ID — also the dbt profile name and the profile's fallback default schema. Staging + intermediate models override it to internal_<DBT_PROJECT_NAME> via +schema: in dbt_project.yml.OUTPUT_PORT_NAME = DATA_PRODUCT_IDCONTRACT_ID = <DATA_PRODUCT_ID>-v1CONTRACT_FILE = <CONTRACT_ID>.odcs.yamlCONTRACT_PATH = models/output_ports/v1/<CONTRACT_FILE>ODPS_FILE = <DATA_PRODUCT_ID>.odps.yamlOUTPUT_PORT_SCHEMA = OP_<TABLE>_V1 (UPPERCASE) — the output-port schema in Snowflake, set literally by the generate_schema_name macro override in macros/get_custom_schema.sql (so it's NOT suffixed with the profile schema). Per dataproduct-builder-dbt's dataproduct-dbt convention the schema is op_<output-port-id>_v<N>; this demo plugin keys off TABLE because it scaffolds a single output port whose id equals the table name (matching the worked example in dataproduct-dbt SKILL.md: op_customer_activity_v1). This is the only public surface; internal staging/intermediate land in INTERNAL_<DBT_PROJECT_NAME>. The contract's servers[].schema and the platform's integration scan target this schema.
Resolve API_HOST from the entropy-data CLI connection: entropy-data connection get -o json → host. If the CLI is not connected, stop and tell the user to run entropy-data connection add <name> --host <host> --api-key <key> first. The demo does not run without a working connection.
Always use the entropy-data CLI for any connection to Entropy Data. Do not use the Entropy Data MCP server.
Step 3 — Scaffold the project
Templates live under ${PLUGIN_ROOT}/skills/dataproduct-bootstrap/templates/. Copy each one into the working directory, substituting placeholders.
| Template | Destination |
|---|
pyproject.toml | pyproject.toml (lists dbt-core, dbt-snowflake, openlineage-dbt, datacontract-cli[snowflake], entropy-data under [dependency-groups].dev so a single uv sync materializes the venv) |
dbt_project.yml | dbt_project.yml |
.gitignore | .gitignore |
README.md | README.md |
profiles.yml.example | profiles.yml.example |
data-product.odps.yaml | <DATA_PRODUCT_ID>.odps.yaml |
openlineage.yml | openlineage.yml |
macros/get_custom_schema.sql | macros/get_custom_schema.sql |
models/input_ports/_models.yml | models/input_ports/_models.yml |
models/staging/_models.yml | models/staging/_models.yml |
models/intermediate/_models.yml | models/intermediate/_models.yml |
models/output_ports/v1/_models.yml | models/output_ports/v1/_models.yml |
models/output_ports/v1/contract.odcs.yaml | <CONTRACT_PATH> |
.github/workflows/data-product.yml | .github/workflows/data-product.yml |
Also create empty analyses/, macros/, seeds/, snapshots/, tests/ directories with a .gitkeep each.
Step 4 — Final report
End with this two-part recap. Use the Status enum: created, already present, skipped.
Part 1 — outcome table.
| Artifact | Status | Details |
|---|
dbt_project.yml | … | profile = <DBT_PROJECT_NAME>, Snowflake adapter |
profiles.yml.example | … | Snowflake |
README.md | … | written |
.gitignore | … | written |
| Model layout | … | models/{input_ports,staging,intermediate,output_ports/v1}/ |
| ODPS | … | <DATA_PRODUCT_ID>.odps.yaml |
| Output-port contract | … | <CONTRACT_PATH> (schema seeded with id + updated_at) |
openlineage.yml | … | transport URL omitted from the file; set at run time via OPENLINEAGE__TRANSPORT__URL env var |
| GitHub Actions workflow | … | .github/workflows/data-product.yml |
Part 2 — next steps. Bullet list:
uv venv && source .venv/bin/activate && uv pip install dbt-core dbt-snowflake openlineage-dbt datacontract-cli entropy-data- This demo assumes
~/.dbt/profiles.yml already has a working Snowflake target. profiles.yml.example is checked in as a reference if you need to recreate the profile elsewhere; otherwise ignore it.
- Set
OPENLINEAGE__TRANSPORT__URL=<your-entropy-data-host> and OPENLINEAGE__TRANSPORT__AUTH__APIKEY=<your-entropy-data-api-key> so dbt-ol run can publish lineage immediately. The committed openlineage.yml omits the URL on purpose so the same repo runs correctly against any deployment (cloud, self-hosted, local) — the URL comes from the env var. The dataproduct-implement skill handles this automatically by deriving both from the active entropy-data connection.
git init && git add . && git commit -m "Initial commit", then push to GitHub.- Set GitHub repository secrets for the workflow:
ENTROPY_DATA_API_KEY, DBT_SNOWFLAKE_ACCOUNT, DBT_SNOWFLAKE_USER, DBT_SNOWFLAKE_PASSWORD, DBT_SNOWFLAKE_ROLE, DBT_SNOWFLAKE_WAREHOUSE.
- Run dataproduct-implement next to generate models from a published data product, run dbt, run dbt + datacontract tests, and ship the first lineage event.
Constraints
- Snowflake-only. This demo plugin intentionally drops platform branching. If the user asks for Databricks/BigQuery/Postgres, point them at the full builder at https://github.com/entropy-data/dataproduct-builder-dbt.
- Greenfield only. If
dbt_project.yml exists, stop.
- Do not run
dbt init. Use the templates here.
- Do not commit secrets.
profiles.yml is in .gitignore; only profiles.yml.example is checked in.
- Do not run
git init, git commit, or git push — surface them as next steps.
- Idempotent. Re-running on an already-populated directory is a no-op (refuse if anything would be overwritten).
