It is odd, now that it has been mentioned. My guess would be that the authors were scared that, if they had a library-related hook, the playtesters would ask for stats for the library...
Ceoris has some. Red Library. Black Library.
There are library hooks, and Durenmar as you say has both hooks and mechanics, and there are occasional mentions of libraries, but in general Ars Magica has strongly moved away from actual texts.
When my survey of covenants is further along, I'll have a better idea of the state of text deposits in the Order. Currently this discussion is as vague as the subject itself.
Although I fully agree on the decission of leaving covenant stats out of the Tribunal books, including libraries, there is one tiny little exception that I do think would be helpful to include, and would only take up one word per covenant out of the wordcount. That one thing is the season of the covenant. Lacking any other stat (which I think is a good idea), the season would be a helpful guideline for Storyguides whenever in need to stat out some part of a covenant, and even more so when having to improvise it. That's the only thing I would really add to covenant descriptions, stats-wise, mostly because it still doesn't nail down a power level or any other saga-specific bit, since each SG could still fine-tune e.g. "autumn" to their saga's current power level.
Still, I do admit it's not really very hard to deduce seasons. But IMHO it would indeed help in judging a baseline for any given NPC covenant, and also would help in setting up hermetic politics games.
On the subject of stats for magi in the various tribunal books...
One thing the tribunal books often do provide is the age of each magus. From here it's fairly easy to extrapolate a hermetic age, and from that you can work out an idea of the potency of that magus.
A bit of a tangent, but I ended up putting together an xp-spend algorithm based on hermetic age which would spit out some arts scores based on whether the magus was a specialist, generalist or balanced magus. It was pretty bucket-value stuff, but it was a close enough approximation that it gave the SG some starting-point numbers.
It meant when the PCs went up against Urgen of Bjornaer he very quickly got stats enough for a conflict to play out.
On the subject of covenant descriptions...
I'm actually happy with what's in the tribunal books. I like that they don't get any more detailing, it lets me as SG decide the role I want the covenant to play in my saga. The only covenants that feel bound in terms of seasons tend to be the Domus Magna - which typically feel like autumn covenants.
Any chance you could share that in some form? Would be great to see.
I understand where you come from on this one, and I mostly agree with you, though I would still think getting covenant seasons wouldn't shoehorn me into any specific plot or role for them. I mean, we do get what they feel about each other and what their relationships are, and those are good things.
I picked that quote just to try to make my point clearer in comparission.
For me, providing covenant seasons would be more or less the same as providing the ages of magi. A quick way to gauge the potency of the covenant, without binding any particular stat.
I second that. 
I agree with the OP. For my style of game libraries are invaluable.
However Tolosa Paratge has a magnificent library focusing on Occitan literature but with many key early texts of the Order. It is definitely not average - it is not as good as Durenmar but better than almost any other covenant.
Generally Provencal libraries are far above average on mundane books and slightly above average in Hermetic ones. The local culture emphasises books.
The algorithm I used is so super-simple it's actually written in wiki-code, which for the not-so-faint-of-heart can be found here:
rpgwiki.geekbraingames.com/rpgwi ... late:Magus
I found that for the younger magi, this worked fine. As the PCs started going past hermetic age 50, the NPCs started to show a slight lead in terms of arts scores. But slight; we're talking visible score differences only past 100. So if I were going to do this again in the future, I'd probably look at making the xp spend per year a curve by age, rather than linear.
It isn't supposed to be simulationist-accurate; it's supposed to give the SG some starting numbers. For things like affinity and puissant, we simply added 3 to the relevant score for each one.
Side note: this wiki is from a now-finished saga. Archiving this one away and making some space for later is on my to-do list.
I've turned your wiki code into an Excel file, but only the part that makes the Arts estimates:
magi-art-scores-estimates-v1.xlsx
While translating your code, I noticed that generalists get more xp than balanced magi, and those in turn get more than specialists, both starting and in yearly increases. More precisely, specialists start with 120xp in Arts, while balanced magi start with 140xp, and generalists start with 150xp. Then, specialists get 23.5xp each year, while balanced magi get 25xp each year and generalists get 25.5xp each year. Is that intentional? I'm asking, because I was expecting each range to get the same starting points (120xp), and the same amount of xp per year (with 25xp looking more or less reasonable).
I'm not able to answer the intent question, but the end result seems reasonable to me - lower level books tend to have higher qualities, which will advantage the generalists.
Yeah, I thought something similar, which is why I didn't just think it was a calculation mistake. But I thought I might just ask. 
I wonder if a more abstract system of books and libraries might be useful for the game. Library detailing can be very time consuming and there are hundreds of things that 'should' be in a 200+ year old covenant library that don't contribute a lot to its value (lab texts on obscure stuff, etc.).
What if the system used for laboratory customization were applied to libraries? You have a 'Size" of a library - call it 'breadth' or 'scope' and you can have 'foci' or 'deficiencies' in the library as well, that modify the 'size' in certain areas. You can have 'trade' to determine how easily your library can acquire desired lab texts, and the actual books that players write can add to 'Size' or improve 'foci'. An entire library could be described in a few short paragraphs in such a manner.
It got tuned based on watching PC development.
I honestly couldn't tell you the exact reasons for the number disparity. There was a reason, but I wrote that some time ago now and haven't opened that box in a while. 
It served its purpose for our saga, and that meant it did the job it was written for.
For anyone who wants to take it and run with it / change it / whatever, go nuts!