I believe the plan is for this to not run on the cluster, rather the parsing will be done somewhere else and the csv will be uploaded and imported.

If someone can confirm that's the case, then we can close this now. Otherwise, the code isn't hostile, but we need to have a defined plan for how it's going to run in production to determine if there are any missing controls.

We're fine either way for now.

Plan from Nik is, "Lets assume for the next few months this'll be manually run and uploaded to the cluster. Or run in labs where the dumps are right there for reading."

If that changes, we'll want to evaluate that process separately.