Context
The jstests framework coverage seems to be quite fragile given the dynamic nature of the javascript language.
The majority of the coverage is given by the core tests which are tested against several settings/events that are defined by every suite.

Problem
We currently don't have a way to actually prove the suites are doing the background work expected.
I managed to gather some tickets that currently prove the problem:

• SERVER-97727: the hook seems to implicitly shard only timeseries collection

• SERVER-118004: FSM test never implicitly sharded the collections
• SERVER-115387: step-downs where never triggered in some "step down suites" for 16 months

We can't claim to have a certain level of coverage if we don't test the suites first.
This ticket propose to investigate a way to prove suite are doing their job.