[FAQ] Why does a generic AI assistant generate invalid Kestra plugin properties?

[c-ibarra]

Course

llm-zoomcamp

Question

Why does a generic AI assistant (e.g. ChatGPT) generate Kestra flow YAML with properties that don't exist?

Answer

A generic assistant isn't grounded in Kestra's current plugin documentation, so it invents plausible-looking property names instead of the real ones.

Concrete example: for io.kestra.plugin.gcp.gcs.Upload, a generic assistant may invent bucket:/name: properties — the real (and only) destination property is to: "gs://bucket/path". For io.kestra.plugin.gcp.bigquery.LoadFromGcs, it may invent separate projectId:/dataset:/table: properties — the real one is a single destinationTable: "project.dataset.table" string.

Always cross-check generated YAML against the official plugin docs (https://kestra.io/plugins/plugin-gcp), or use Kestra's built-in AI Copilot, which is grounded in the current plugin schema for your running version.

Checklist

• I have searched existing FAQs and this question is not already answered
• The answer provides accurate, helpful information
• I have included any relevant code examples or links

[github-actions]

✅ FAQ NEW proposal created in PR #298

Please review and approve the changes.