I don't think there's a case for [adj-f]. The etymology is not really appropriate here.

@@ -9 +8,0 @@
-<pos>&adj-f;</pos>
@@ -13 +12 @@
-<s_inf>attributive form of き, auxiliary verb indicating directly experienced or factual past; roughly equivalent to modern 〜た</s_inf>
+<s_inf>equiv. of modern 〜た</s_inf>

daijr, daijs, meikyo

Added [adj-f] as し is 連体形. Edited notes to include more information with the intent of avoiding potential confusion.

I would also like to include how it connects to verbs but き has irregular coupling methods to both サ変 and カ変 verbs that I fear would leave the entry too long and not including them might make it seem like it does not do so at all; Apologies if this is the wrong place to ask but has something such as a section that includes the inflection paradigms of irregular verbs as well as notes on more specific irregularities such as these ever been considered? Items that would benefit from this would be honorifics such as なさる and おっしゃる, as well as ます and defective helpers such as ぬ・ず and う・よう, and also archaisms such as べし・べき; pretty much anything with some irregularity, whether true or perceived, that the learner cannot foresee and consequently tends to struggle to identify.

@@ -8 +8,2 @@
-<pos>&aux;</pos>
+<pos>&aux-v;</pos>
+<pos>&adj-f;</pos>
@@ -12 +13 @@
-<s_inf>literary form of 〜た</s_inf>
+<s_inf>attributive form of き, auxiliary verb indicating directly experienced or factual past; roughly equivalent to modern 〜た</s_inf>

I think an xref is enough.

@@ -8,0 +9,2 @@
+<xref type="see" seq="2178970">選ばれし者・えらばれしもの</xref>
+<xref type="see" seq="2654250">た・1</xref>
@@ -10 +12 @@
-<s_inf>literary form of -た-, e.g. 選ばれし者</s_inf>
+<s_inf>literary form of 〜た</s_inf>

format based on ざる