- sankoku has 端たない as a headword (wrapped in its '仮名書き可' parentheses)
- shinmeikai has this word as a subentry in its entry for はした【端た】
- obunsha has 端ない as a headword
- iwakoku states directly that this is 「端」の形容詞化。

None of these refs say it is ateji or a special reading / jukujikun.

Google N-gram Corpus Counts
╭─ーーーーー─┬────────┬───────╮
│ 端ない　　 │     95 │  0.2% │ - [iK] to [sK]
│ 端たない　 │    120 │  0.3% │ - [iK] to [rK]
│ はしたない │ 38,450 │ 99.4% │
╰─ーーーーー─┴────────┴───────╯

@@ -5,2 +5,3 @@
-<keb>端ない</keb>
-<ke_inf>&iK;</ke_inf>
+<keb>端たない</keb>
+<ke_inf>&rK;</ke_inf>
+<ke_pri>ichi1</ke_pri>
@@ -9,3 +10,2 @@
-<keb>端たない</keb>
-<ke_inf>&iK;</ke_inf>
-<ke_pri>ichi1</ke_pri>
+<keb>端ない</keb>
+<ke_inf>&sK;</ke_inf>
@@ -20 +19,0 @@
-<s_inf>poss. ateji</s_inf>

Since the 国語s are quiet on it being an ateji, I think I'll put it in a note, and close it.

@@ -19,0 +20 @@
+<s_inf>poss. ateji</s_inf>

Sorry, I meant [iK], not [oK]. Definitely irregular, not obsolete.
I can find quite a few recent examples of "端ない" on Twitter.

@@ -6 +6 @@
-<ke_inf>&oK;</ke_inf>
+<ke_inf>&iK;</ke_inf>
@@ -10 +10 @@
-<ke_inf>&oK;</ke_inf>
+<ke_inf>&iK;</ke_inf>

is there any evidence that this is "obsolete"?  (i.e., is there any evidence it used to be a thing in the past but is no longer?  the uses i'm seeing online don't seem to come from old sources)  i think ateji makes more sense
--
also, 端たない seems to be more common than 端ない, since 端ない looks to get a lot more false positives.
--
overall i get <200 "real" google hits for both headwords combined (many/most of which don't seem to be real), which raises the question of whether it's worth recording at all, or if it's just a henkan error and should be iK