>Problem is that unknown unknows are not actionable.

How about carefully testing something new, especially something that one ingests? As far as I know no such testing is done, if it's done, it's done with the goal of gaming the safety standards that may be expected of the product. You cannot fix an ignorant, corrupt society.

How extensive safety testing of a hammer should be?

Should there be double blind clinical trials for it’s health impact when caried daily on a tool belt?

For most people, answer will differ depending on what the hammer is made of (iron vs uranium vs radium).

My point here would be that a) we must choose appropriate set of tests as they cost in money/time/opportunities, b) the choice of tests must be influenced by what we know and/or suspect.

If unknown is unknown, then we don’t know that we need a test for it.