As a preface - it is a recurring topic, fragments of the relevant discussion can be seen here on Meta. Even in my relatively short time here, I have seen a few persistent strains of "is X plagiarism", with the most common confusion seemingly coming from the stance on paraphrasing. A few academic integrity resources such as this MIT one shed some light on the issue, as they dive straight into "borrowing exact words without attribution is plagiarism", which - in my opinion - fails to establish a solid foundation and convey underlying concerns.
Should we have a canonical question covering all (well, to the extent it could be done) things plagiarism?
Usually, I find myself conflicted because of not having a good duplicate target for VTC and end up writing an answer. And then try to make it not just a collection of links - that almost screams "canonical question".
EDIT: (apologies for being late) As cag51 points out, the overall issue about plagiarism and authorship is broad, and the corresponding tag has plenty of questions along the lines of "someone stole my work or is suggesting that I steal others' work, what do I do?". While the linked question covers many of the bases for "is X plagiarism?", I still feel it is not quite there as a duplicate target. Mostly because it assumes a bit different perspective from what askers typically have in mind. I believe that a more FAQ-like answer or a collection thereof could be helpful, but our regular format does not allow that; questions either end up being too broad or violate one question per post rule.
There is also an argument to be had that providing these detailed answers is outside of SE's scope, as they are covered elsewhere on the Internet, but people keep asking them here. And we keep answering them here; they are still legitimate. But it seems - at least, to me - the redundancy is becoming quite high.