At my previous employer, I was the QA Supervisor, and the QA Manager and I devised a simple annual proficiency test for our QA techs wherein they were required to perform the laboratory testing processes/techniques that were required of them (testing for water activity, moisture, particle size, density, etc.) against a standard.
The QA Manager established a standard for each element they would be tested on, and would provide the necessary "raw materials" for each test, so that there was no variability between tests (i.e. everyone used material from the same sample bag to ensure all things were equal, aside from personal proficiency at the task at hand). If the tech's results were within the established specification, we simply signed off on their testing for the year. If not, the tech was re-trained on the procedure, and then could attempt the test again.
I don't know what the QA Manager had planned if the tech couldn't pass on the second try, because we never got to that point - everyone passed on their first or second attempt.
To my knowledge, the SQF auditor had no problem with our approach to this.