Agent Skill

@jrmc/etl ships an agent skill that teaches AI coding agents (Claude Code, Cursor, Codex, etc.) how to build pipelines with this package: the etl.run API, the three accepted input shapes and when to use each, component scaffolding conventions, ESM pitfalls, and complete recipes (XLSX to database, database to CSV).

Installation

From your project root:

npx skills add batosai/etl

The CLI detects the etl-pipeline skill in this repository and installs it for your agent.

What it provides

Once installed, your agent automatically loads the skill when you ask it to import, export, or migrate data with @jrmc/etl. It then knows how to:

• choose between inline functions, lazy-import classes, or the [LazyImport, options] tuple
• scaffold Source / Transform / Destination classes with the right interfaces and default exports
• respect ESM constraints ("type": "module", .js extensions in relative imports)
• pick a consistent granularity (item by item vs batch) across the three components

Example prompts

Import users.xlsx into my database with @jrmc/etl

Create an ETL destination that writes rows to a CSV file

The skill source lives in the skills/etl-pipeline directory of the repository.