All products

Edit Metabase report files with AI

Work on exported YAML for dashboards, questions or collections while keeping the identifiers and parent relationships that Metabase uses on import.

You can edit Metabase report files with AI when analytics content lives in version control and colleagues need to review changes before they reach an instance. The metabase-representation-format skill helps the agent identify each entity, read its current format and preserve the relationships around it.

Moving a collection illustrates why this matters. Shifting a directory on disk looks like a completed move in a file browser. Metabase derives the collection’s parent from the identifier inside its YAML, so the edit needs to express that relationship explicitly.

What it does

  • Identify the file that defines a saved question, dashboard or collection.
  • Follow parent and ownership references between entities.
  • Preserve stable entity identifiers when editing existing content.
  • Distinguish an importable top-level directory from a convenient local folder.
  • Use the maker’s schema checker to inspect a revised file.

Install and use

  1. Install the skill and provide an authorised export. Node and the official Metabase representations package supply the format specification and schema validator.
  2. Ask “Move this collection under quarterly reporting. Show the parent identifier and file changes before importing anything.”
  3. Review the diff and validate file structure. Check references separately if you also need proof that a target entity exists.

FAQ

Source

Made by Metabase. Original skill: metabase-representation-format, published on GitHub under the MIT licence.

View on GitHub