Some mention has been made in this thread of how first-party packages should properly be referred to and what disclosures are necessary. Since this topic doesn't seem to have been addressed in any detail its own right, I would like to record my thoughts on the subject and raise it for further discussion.
Current policy would appear to be that, if one has written a package (or similar) not specifically for the purpose of answering a question on this site and wishes to use it in an answer, it is necessary to state quite explicitly that one is the author by making some kind of disclosure statement. Additionally, it is not considered acceptable if all of a user's answers are of this type.
My view, however, is that it doesn't matter if someone exclusively posts answers featuring their own work. Indeed, in a very real sense, all good answers should do so. Problems only arise if these answers become unreasonably repetitive and uninteresting, or if they are so tangentially relevant to the question as to strain credibility. I also think that including a disclosure statement into every single answer is not really necessary, especially if this information is already noted in one's profile, or if the package in question is freely available without the requirement for payment or registration of any kind. Disclosure statements, I would argue, are required only where a conflict of interest might otherwise be perceived.
As an example, it seems somewhat ridiculous to me that Szabolcs should feel obliged to disclaim his (co-)authorship in answers mentioning the MATLink package, since he and the other author wrote this package at least partly as a service to the community and have released it free of charge under the extremely liberal MIT licence. It is hence obvious that no tangible benefit accrues to them when a user uses this package, and therefore that no conflict of interest is possible. This is not to say that recording one's authorship doesn't constitute useful information, but merely that a disclaimer should not be necessary in this situation.
Perhaps the current policy is sensible simply to avoid any possibility of doubt regardless of the circumstances. However, I find it extremely incongruous when answers contain such disclaimers, as if having contributed to the library of third-party Mathematica packages is something to be ashamed of.
Please indicate your agreement or disagreement with my point of view by upvoting or downvoting this question as appropriate.