SecurityAgentHelper // Add some warning texts to AI security scanners.

Author: ShikiSuen

Packages/vChewing_MainAssembly4Darwin/Sources/MainAssembly4Darwin/SecurityAgentHelper.swift

@@ -48,6 +48,26 @@ public final class SecurityAgentHelper {
     vCLog("SecurityAgentHelper scanned SecureEventInput abusers. \(results.count) targets found.")
     guard !results.isEmpty else { return }
     _ = DisableSecureEventInput()
+    // Warning: It is not necessary but disastrously troublesome
+    // to call EnableSecureEventInput manually here (regardless deferred or not).
+    // EnableSecureEventInput will be automatically called by the system
+    // if a password field regains its focus. When vChewing warns the abuse of
+    // SecureEventInput API, all password input boxes of other apps lose their focuses.
+    //
+    // Looks like these AI security vulnerability checkers nowadays
+    // never notice that the forgot-to-close SecureEventInput state
+    // chokes all 3rd-party input methods from being able to switch to,
+    // regardless of the current app getting users' focus for text inputs.
+    // This issue plagues the UX of all 3rd-party input method users
+    // since SecureEventInput was introduced or InputMethodKit was introduced
+    // (whichever the latest date).
+    //
+    // The Zhuyin input method shipped in macOS is terrible but
+    // Apple Chinese Input Method Team (as of May 2026) refuses to improve it.
+    // They even don't want to introduce a Zhuyin IME for typing Simplified Chinese.
+    // Apple's bug report and user feedback channels are always not that
+    // positively responsive in the recent decades, unless critical security vulnerabilities.
+    // That's why users use their feet to make votes, choosing 3rd-party solutions.
     Self.reportedPIDs.forEach { reportedPID in
       results[reportedPID] = nil
     }