# Create Bauplan Data Pipelines

Bauplan projects need clear DAG structure, branch safety, and consistent SQL or Python model patterns. This skill guides Claude, Codex, and Claude Code through pipeline files, validation, and safe run workflows.

## Install

```bash
npx skillstore add bauplanlabs/creating-bauplan-pipelines
```

## Metadata

- Status: approved
- Slug: bauplanlabs-creating-bauplan-pipelines
- Skillstore revision: r1
- Version status: missing
- Tree hash: 443a44e5788041cce5ba86b19df69514211b480f13352a983cb4702c0e564c5b
- Author: BauplanLabs
- GitHub username: BauplanLabs
- License: MIT
- Repository: https://github.com/BauplanLabs/bauplan-mcp-server/tree/main/skills/new-pipeline
- Ref: 5ab1d37e83436c6eef84de88573e142b94f2a4ad
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: allowed
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: scripts, external\_commands
- Quality score: 70
- Public page: https://skillstore.pages.dev/skills/bauplanlabs-creating-bauplan-pipelines
- Manifest: https://skillstore.pages.dev/api/skills/bauplanlabs-creating-bauplan-pipelines/manifest

## Capabilities

- Creates Bauplan project structure with a project configuration and model files.
- Guides SQL first-node models that read from existing lakehouse tables.
- Guides Python transformation models with Bauplan decorators and pinned package versions.
- Supports materialization choices for final outputs, including REPLACE and APPEND.
- Adds data quality expectations and output column validation patterns.
- Provides dry-run and strict-mode workflow steps for pipeline checks.

## Use Cases

- Start a New Analytics Pipeline: Create a Bauplan project with source nodes, transformations, materialized outputs, and a dry-run workflow.
- Convert Business Logic into Models: Turn transformation requirements into SQL source models and Python downstream models with clear DAG dependencies.
- Add Data Quality Checks: Define expectations, output schemas, and strict-mode checks before publishing pipeline results.

## Prompt Templates

### Create a Basic Pipeline

```
Create a Bauplan pipeline for [goal] using source table [table]. Ask me for missing branch, output, and materialization details before writing files.
```

### Add Python Transformations

```
Build a Bauplan DAG for [business logic]. Use SQL only for first source reads, then use Python models with declared output columns.
```

### Design Multi-Input Models

```
Create a multi-input Bauplan model that joins [table A] and [table B]. Use I/O pushdown, docstrings, and materialized final outputs.
```

### Harden and Validate a Pipeline

```
Review this Bauplan pipeline for branch safety, materialization choices, strict mode, expectations, and schema validation. Suggest fixes before any run.
```

## Limitations

- Requires the Bauplan CLI and user credentials to be configured outside the skill.
- Assumes source tables already exist in the Bauplan lakehouse.
- Does not validate live schemas unless the assistant runs approved Bauplan CLI commands.
- Generated transformations still need review against business rules and data contracts.

## Best Practices

- Use a development branch and run a dry run before any full pipeline run.
- Declare output columns and use I/O pushdown for every model when possible.
- Confirm materialization strategy for each final table before running the pipeline.

## Anti Patterns

- Running pipeline commands on the main branch.
- Using SQL models for downstream transformations after the first lakehouse read.
- Leaving branch names, table names, or SQL preview queries unchecked before execution.

## Security Audit

- Audited at: 2026-07-05T04:44:44.286\+00:00
- Summary: The static findings are false positives from markdown backticks, Python import examples, table samples, and assertion text. I found one contextual medium risk: user-provided branch and table identifiers are inserted into Bauplan CLI and SQL preview commands without validation guidance.

## Stats

- Views: 281
- Downloads: 8
- Favorites: 0
- Popularity score: 0