I think you should reevaluate your very concept of "official", high detailed source for all country borders worldwide.
For a really "official" source, you would have to check 200 +/- official sources of all +/- 200 existing countries wordlwide. Borders are defined on a bilateral base. The actual delimitation takes place between two neighboring countries. For that reason, there cannot be an "official" registry of all worldwide borders, but at best +/- 200 "national" datasets. There is no international body (like UNO) registering the official definitions of all conturies, this is strictly a matter of bilateral agreements - and in many cases still of custom and practice (customary law) - at least for a precision of less then 100 meters and in more remote areas (deserts, swamps, forests, mountains, lakes etc.).
So every insititution that collects these data (like OSM) is not an official source any more by definition. If you really are to get official definitions, you can get them only at the national level.
In many cases, not even that does help much. Neighboring countries often have diverging definitions of what belongs to them and what to their neighbors. Georgia and Russia will have different definitions on their borders - not to speak about Crimea.
There are many other cases: the delimitation of the borders between Uzbekistan, Kyrgyzstan and Tajikistan in the Ferghana valley have been an issue for decades now. We not only speak about clear cases of contested territories, but also of generelly undisputed portions of the border that are not delimited - thus no unanimously shared definition in the terrain (like boundary stone) does exist. Only if both parts sign an agreement that fixes a border in all it's details (most often still based on real boundary stones put in the terrain by a bilateral commission) can we speak about an "official" border.
As border zones tend to be a delicate issue for security concerns, many states ar not keen to publish official documentations about this delimitation. And a map in always all cases is a derivative product that documents one kind or another of border-definitions, but it is not the (legally binding) definition in itself - so to get the "official" source would meen going to the archives (if you have access) to access protocols on the delimitation and demarcation (like the ones of the Kyrgyz-Uzbek borders mentioned here).
And even today, not even such protocols exist. Many borders are not defined in such a precise way, but are still based on older principles like "in the middle of the river" - riverbeds change, however. There are sections of the border between Switzerland and Italy that are defined as "the ridge of a glacier". With climate change and melting glaciers, the border changes, too. As well, there is no agreed definition of the border between Switzerland and Germany in the Lake of Constance - as there is no need for it:
there is no legally binding agreement as to where the borders lie between the three countries. See: https://en.wikipedia.org/wiki/Lake_Constance#International_borders
See also this statement by Swisstopo (Swiss Federal Office of Topography):
Our boundaries can change: Borders are commonly regarded as fixes, but depending on the terrain they are subject to climatological changes and natural phenomena and may thus change their course.
So even in the middle of Europe with it's sophisticated administrative institutions, high precision cartographic technologies, need for clear definitions in a denseily populated area and centuries-long tradition of border delimitation, borders to this day sometimes are still not defined "on the meter". This is even more true for many parts of the world.
So I'm not sure if the very concept you're after does makes sense. With no dataset you'll get a guarantee that it is correct or generally accepted. The more "official" and detailed the data gets, the more you'll get an illusion of precision and accuracy that in fact does not exist.