I'm not sure the question is an improvement nor that the answer is as simple as you'd like.
In a world of black and white and where engineers have autonomous decision taking authority then the answer may be different to world filled with greys.
For example, competence might have to include being able (having the necessary judgement) to only do as much as is necessary or required to deliver the required result.
This goes to the old saw:
"Rules are for the observance of fools and the guidance of wise men."
Incompetence then becomes a failure either to follow the rules or make the correct judgement as to what needs and does not need to be done.
In the other thread, following a procedure documented by others reveals a possible vulnerability which is the question of the verification of the original procedure as being sufficient or flawed either when it was written or as subsequently affected by changed circumstances/knowledge.
The real world problems are the bean counters, the people who look over our shoulders and who use documented procedures as the metric by which they judge our performance.
Doing less than the procedure calls for may cost less and may not attract attention from the bean counters.
But doing more than the procedure calls for will create and extra cost to be assigned which certainly will attract attention.
If there is a real issue with the procedure then the procedure needs changing.
That is never as easy as it seems.
This means that there is another dimension to the problem.
The low profile approach to such problems is to accept the procedure and do no more than is required unless there are compelling reasons to instigate a change to the procedure.
If the procedure doesn't have any significant "risks" attached then the judgement may be that in most cases the procedure is adequate. The "risk" being a non catastrophic failure at some point which will result in a high cost outcome. Then the procedure may be revisited so that the risk of the high cost outcome can be avoided.
In the modern world of engineering, companies try to subsist on fewer engineers. Engineers who would like to choose solution "a" which costs more are often well advised to accept "b" which costs less (capital cost) but which may well cost more over time (cost of ownership) whatever their natural inclination.
These days of "single source" suppliers give evidence that it is the bean counters and buyers who get to choose, and far less the engineer.
Or to take another example.
A company provides turnkey solutions, complete processing plants. Bidding costs money. SO they can either bid each new job starting with a blank sheet of paper or they can simply take the previous bid, include factors for currency fluctuations etc. and make allowances for variances.
In one particular example the engineers knew that one instrument specified in the bid never worked and would be replaced during commissioning with a more suitable instrument. But they were not allowed to change the original specification to delete the wrong instrument and insert the correct one because it altered the base cost. With enough such changes it would affect the validity of the bidding mechanism. It was enough that the variances (changing instruments after the event etc) were accounted for in the bid evaluation procedure.
Not everything is as logical as we'd like and not everything we object to is necessarily wrong, it is just we don't necessarily understand the external logic (i.e the other factors besides engineering).
So the question of competence is complicated by other factors, not least, the question of survival.
Competence is then a case of recognising the true extent of the objectives (which includes an understanding of the implications of costs/risks etc.)and acting accordingly.
JMW