If two reviewers grading the same call disagree, the criterion is broken — not the reviewer.
Rubrics start as a conversation with an operator: what do you check before you call this call handled? The first draft is always a list of instincts, and instincts do not survive being handed to a second person.
Every criterion then gets rewritten until two people grading the same call land on the same score. Anything that still splits reviewers goes back to the operator or gets cut. Persistent disagreement is treated as a schema problem, which is why we publish agreement per batch rather than an overall quality claim.