The terms for researchers examining potential violations of research integrity are bad (data police, data thug).
What is an alternative term that:
* doesn't imply anything about the intentions of the people who did the work being investigated (deliberate or accidental)
* doesn't imply the researcher has any power over the people/work being investigated
* does indicate the work is scholarly, and should be valued
* does indicate that "self-appointment" to investigate is appropriate & ordinary