diff --git a/bin/omarchy-update-perform b/bin/omarchy-update-perform
deleted file mode 100755
index 3520a138..00000000
--- a/bin/omarchy-update-perform
+++ /dev/null
@@ -1,7 +0,0 @@
-#!/bin/bash
-
-# omarchy:summary=Compatibility wrapper for omarchy-update -y.
-# omarchy:hidden=true
-# omarchy:requires-sudo=true
-
-exec omarchy-update -y
diff --git a/docs/update-process.md b/docs/update-process.md
index bcc0c86e..0810fa68 100644
--- a/docs/update-process.md
+++ b/docs/update-process.md
@@ -247,7 +247,6 @@ scripts.
 | `omarchy-update-lock` | Hidden command wrapper that holds the per-user update lock while its child runs. | **Keep internal/hidden.** Isolates update concurrency and lock descriptor handling. |
 | `omarchy-update-stay-awake` | Hidden helper that starts or stops update-owned sleep and idle inhibition, restoring only the state it changed. | **Keep internal/hidden.** Keeps inhibitor ownership and cleanup together. |
 | `omarchy-update-status` | Hidden helper that refreshes or clears the shell update indicator after rechecking available updates. | **Keep internal/hidden.** Keeps shell status synchronization out of the main pipeline. |
-| `omarchy-update-perform` | Hidden compatibility wrapper for `omarchy-update -y`. | **Temporary.** Keep only for old callers; new code should call `omarchy-update` directly. |
 | `omarchy-update-confirm` | Gum confirmation copy for `omarchy update`. | **Question.** Could be inlined into `omarchy-update`; separate file only helps keep copy isolated. |
 | `omarchy-update-dev` | Fast-forwards the active dev-linked checkout from its configured upstream; no-ops for package-backed installs. | **Keep.** Runs before package updates so a checkout conflict stops the update before system mutation. |
 | `omarchy-update-keyring` | Ensures Omarchy keyring and Arch keyring are current before the main transaction. | **Keep, but review.** It uses targeted `pacman -Sy` for keyring bootstrapping; acceptable for this special case but should remain tightly scoped. |
@@ -280,8 +279,6 @@ scripts.
 
 3. **Update pipeline ownership**
    - `omarchy-update` owns the full update pipeline now.
-   - `omarchy-update-perform` is only a hidden compatibility wrapper for
-     `omarchy-update -y`.
 
 4. **Mise remains in the blessed update path**
    - `omarchy-update-mise` intentionally runs as part of `omarchy update`.
diff --git a/test/shell.d/update-lock-test.sh b/test/shell.d/update-lock-test.sh
index 64699e1c..55fcd92e 100644
--- a/test/shell.d/update-lock-test.sh
+++ b/test/shell.d/update-lock-test.sh
@@ -75,32 +75,6 @@ grep -q "already running" "$test_tmp/update-second.out" || fail "second omarchy-
 [[ ! -f $test_tmp/update-second-snapshot-started ]] || fail "second omarchy-update did not snapshot while lock was held"
 pass "omarchy-update prevents overlapping top-level updates"
 
-# omarchy-update-perform is now only a compatibility wrapper around
-# omarchy-update -y, but it should still respect the same update lock.
-perform_marker="$test_tmp/perform-started"
-write_stub omarchy-update-keyring 'echo started >"$TEST_MARKER"; sleep 2; exit 0'
-
-TEST_MARKER="$perform_marker" run_with_lock_env "$ROOT/bin/omarchy-update-perform" >"$test_tmp/perform-first.out" 2>&1 &
-perform_pid=$!
-
-for _ in {1..50}; do
-  [[ -f $perform_marker ]] && break
-  sleep 0.05
-done
-[[ -f $perform_marker ]] || fail "first omarchy-update-perform delegated to update under lock"
-
-set +e
-TEST_MARKER="$test_tmp/perform-second-started" run_with_lock_env "$ROOT/bin/omarchy-update-perform" >"$test_tmp/perform-second.out" 2>&1
-perform_second_status=$?
-set -e
-
-wait "$perform_pid"
-
-[[ $perform_second_status -ne 0 ]] || fail "second omarchy-update-perform exits non-zero while update lock is held"
-grep -q "already running" "$test_tmp/perform-second.out" || fail "second omarchy-update-perform reports held update lock"
-[[ ! -f $test_tmp/perform-second-started ]] || fail "second omarchy-update-perform did not snapshot while lock was held"
-pass "omarchy-update-perform compatibility wrapper respects update lock"
-
 # The sleep inhibitor deliberately outlives the step that starts it, so it must
 # not inherit the update lock. An update killed before restore_update_inhibitors
 # would otherwise leave the inhibitor holding the flock forever, blocking every