The Russian lessons are a trat of Pije, so, of course, similar. Not expanding much here.
The items you have picked up are interesting:
a comma with 'li' is just weird, I can't imagine anyone (even Sonja) doing it deliberately. I wouldn't write a rule for it (but, then, I am not sure about writing rules for punctuation anyhow). The commas with 'ala' are not much better but more common and more scattered.
Sonja does use unofficial words after 'li' (actually, just one occurrence, but that is the most we get for most things). It violates so many of her own rules and common sense that I don't see the point in legitimizing it in your rules. Of course, I can' really hope anyone will get it really right, but I can sorta live with that.
'o' requires a person, but what is a person is contextual. Here, apparently, a bug makes it in. (Circular, since one of the tests of person hood is being addressed.)
'mute seme' for "how many?" is standard and old. Savants just haven't noticed or didn't bother to mention it.
The headline example you give is a sentence, not an object phrase (?) but you do need to deal with Frag as an utterance type (basically any well-formed sentence part, so NP, Pred).
'la' at the end is novel and probably wrong (imagine 'la S' there). It probably arises from a confusion , since many 'la' phrases can be shifted to prep phrases at the end 'tenpo kama la' <> 'lon tenpo kama', etc.
The 'taso' you cite is before a sentence, not before a 'la' phrases (though the sentence does begin with a 'la' phrase)
'len' as a transitive verb meaning "dress" is standard (and by the rules) and old, as is 'supo' as an adjective meaning "horizontal, flat, etc."
'esun' as a verb is the original 'esun', meaning "to barter, swap". The "store" use was originally derivative ('tomo esun'). There were problem getting from "swap" to either "buy" or "sell" , which the new definition has not eliminated (but rather complicated).
[I note in passing that, had you checked my stuff, you would have had much of this already.]