CSV validation usually starts with rules: required fields, types, ranges, allowed

values, and uniqueness.

But if the validation output will be attached to an issue or sent to a teammate,

there is an earlier design question:

Should the report repeat the value that failed?