Maybe we can agree that "harmful" means something counter to the primary stakeholders. I think that's the primary intent. Of course it doesn't prevent virus writers, since their primary stakeholders are the scum who hire them. It's a hard point to make. It has the same problems in the hippocratic oath for doctors "do no harm". It's well meant, but fuzzy.
Defective in behaviour likely means it doesn't fulfill it's requirements. Defects in structure probably means it fulfills its requirements but has problems at the code level, possibly even failing certain non-functional requirements.
Note, I agree with the other points. I wish everybody cared a lot about the work they do.
In programming, there is always a strong back-and-forth between what people strive for and what they end up with because of reasonable tradeoffs. I find folks like Bob Martin usually concede that this is clearly the case, but they don't always do a good job of acknowledging that reality in the initial statement.
The open/closed principle is an example of a concept that I find is a nice ideal, but none of my work has really been able to get anywhere near complying with it on account of difficulty predicting the future, so I've found it to be a bit unhelpful in that it is rarely described as a good idea that might not come to be in practicality.
Perhaps some of these concepts would be best understood if the room for error was more baked into the initial wording rather than something tacked on when discussing the matter in practical terms.
We're a place where coders share, stay up-to-date and grow their careers.
We strive for transparency and don't collect excess data.