I imagine that conflation of "vocab level" and "kanji level" creates problems.
はし (chopsticks) is an N5 word, but when written as 箸 it is N2 level.
I think it includes alternative kanji forms in the equation too.
I'm not sure if there's a way to get around that as it currently stands.
I don't think it's supposed to be adding them together though, you have specific tags for looking up n-related kanji as well as just general vocab.
kanji #jlpt-n5 and #words #jlpt-n5
The "kanji" and "words" tags seem to just limit which type of entries are displayed by the search, rather that limiting the way an entry is evaluated.
Perhaps the creator envisioned it the way you're wishing it worked and it's just a bug. I guess we'll have to wait for them.
Was looking to create a list of words that I know/don't know so I started to filter stuff by tags, after #words #jlpt-n4 , n3, n2 and n1 have mistakes in them, n3/n2 list words from n4/n5 and the n1 list has words from n5/4/3/2 as well as having verbs in it as well - This should be nouns/na-adjectives/no-adjectives only.
Not sure if the system was designed for this but it works for n5 adjectives: #words #adj-i #jlpt-n5 shows 59 words, n4 shows 19 - whether or not this is correct I don't know but n3 shows 29 results but a bunch of them are from n5.
If you do just a basic #noun search, the order of the words doesn't make sense, you have n3 words on page 100, n2 words on page 50 and the second half of the n5 list appears on page 35.
It seems that the n5 tags are working fine, but beyond that errors start to appear.