The footnote implies that intermediate forms were not written down. Probably everyone involved has now forgotten who contributed what.

Is it meaningful to ask about the problem? In what sense is Pascal's mugging different than Pascal's wager? That it only uses finite numbers? That it uses super-exponential growth and draws attention to Kolmogorov complexity? That the mugger's number comes second, allowing the threat/promise to depend on the probability? And, of course, the iconic difference, the one in the name, is that it is a single person making the claim, not a society, yet this is not a technical difference, but a purely psychological difference, and thus quite ambiguous and difficult to trace. At some level of granularity, every account is the same; at another, every account is different.

Commenting on Yudkowsky's post, Bostrom cites his paper "Infinite Ethics" (section 4.3?). Presumably that is as close as Bostrom got in writing before. Whether you consider it "the same" or an admission of not having prior art depends on what you care about. Moreover, that paper has a bibliography, unlike the mugging dialogue.