diff --git a/doc/todo/command_to_list_keys_for_old_versions_of_files/comment_1_43c11d9467b4d4460aa97bfb88b4896f._comment b/doc/todo/command_to_list_keys_for_old_versions_of_files/comment_1_43c11d9467b4d4460aa97bfb88b4896f._comment
new file mode 100644
index 0000000000..34183374ed
--- /dev/null
+++ b/doc/todo/command_to_list_keys_for_old_versions_of_files/comment_1_43c11d9467b4d4460aa97bfb88b4896f._comment
@@ -0,0 +1,18 @@
+[[!comment format=mdwn
+ username="nobodyinperson"
+ avatar="http://cdn.libravatar.org/avatar/736a41cd4988ede057bae805d000f4f5"
+ subject="Oh yes please 🤩"
+ date="2023-08-22T16:48:58Z"
+ content="""
+`unusedkeys` would be consistently named with `git annex unused`. Would it make sense to teach `git annex unused` that? It could be taught to accept paths as arguments (it doesn't currently, right?) and then only operate on those to find old versions. Then no new command is necessary and people already know `unused`. The workflow would be:
+
+```
+# two-step
+git annex unused file.txt
+git annex move --unused --to there
+
+# new option to only list keys for further scripting
+git annex unused --onlykeys file.txt | xargs ...
+# maybe --batch is a better name 🤷
+```
+"""]]