fix(cua): distinguish disabled paste commands and verify fixture focus

packages/computer-use/cua/native/Sources/CuaNative/AXAccess.swift

@@ -164,14 +164,14 @@ final class AXAccess {
             if try string(node, "AXRole") == "AXMenuItem", try string(node, "AXMenuItemCmdChar")?.lowercased() == "v" {
                 let raw = try attribute(node, "AXMenuItemCmdModifiers")
                 if let raw, CFGetTypeID(raw) == CFNumberGetTypeID(), let number = raw as? NSNumber, number.intValue == 0,
-                   try actions(node).contains("AXPress"), try boolean(node, "AXEnabled") != false {
+                   try actions(node).contains("AXPress") {
                     guard found == nil else { throw NativeFailure(code: "UNSUPPORTED", message: "The application exposes ambiguous Paste commands") }
                     found = node
                 }
             }
             for child in try children(node) { pending.append((child, depth + 1)) }
         }
-        guard let found else { throw NativeFailure(code: "UNSUPPORTED", message: "No enabled standard Paste command is available") }
+        guard let found else { throw NativeFailure(code: "UNSUPPORTED", message: "No standard Paste command is available") }
         try beforeDispatch()
         try perform(found, "AXPress")
     }

packages/computer-use/cua/native/tests/CapabilitiesFixture.swift

@@ -25,7 +25,7 @@ final class CapabilityPad: NSView {
     }
 }
 
-final class CapabilitiesFixture: NSObject, NSApplicationDelegate, NSTextViewDelegate {
+final class CapabilitiesFixture: NSObject, NSApplicationDelegate, NSWindowDelegate, NSTextViewDelegate {
     let statePath: String
     var window: NSWindow!
     var text: NSTextView!
@@ -44,6 +44,7 @@ final class CapabilitiesFixture: NSObject, NSApplicationDelegate, NSTextViewDele
         NSApplication.shared.mainMenu = bar
         window = NSWindow(contentRect: NSRect(x: 140, y: 160, width: 640, height: 500),
                           styleMask: [.titled, .closable], backing: .buffered, defer: false)
+        window.delegate = self
         window.title = "DSH CUA capability checks"
         window.isReleasedWhenClosed = false
         window.animationBehavior = .none
@@ -74,6 +75,10 @@ final class CapabilitiesFixture: NSObject, NSApplicationDelegate, NSTextViewDele
         NSApplication.shared.activate(ignoringOtherApps: true)
         persist()
     }
+    func applicationDidBecomeActive(_ notification: Notification) { persist() }
+    func applicationDidResignActive(_ notification: Notification) { persist() }
+    func windowDidBecomeKey(_ notification: Notification) { persist() }
+    func windowDidResignKey(_ notification: Notification) { persist() }
     func textDidChange(_ notification: Notification) { persist() }
     func textViewDidChangeSelection(_ notification: Notification) { if pad != nil { persist() } }
     func persist() {
@@ -91,6 +96,8 @@ final class CapabilitiesFixture: NSObject, NSApplicationDelegate, NSTextViewDele
         let frame = pad.convert(pad.bounds, to: nil)
         let selection = text.selectedRange()
         let result: [String: Any] = ["pid": ProcessInfo.processInfo.processIdentifier, "windowId": window.windowNumber,
+            "active": NSApplication.shared.isActive, "keyWindow": window.isKeyWindow,
+            "textFocused": window.firstResponder === text,
             "text": text.string, "selection": [selection.location, selection.length], "runs": runs, "events": events,
             "pad": [frame.minX, window.frame.height - frame.maxY, frame.width, frame.height]]
         try! JSONSerialization.data(withJSONObject: result).write(to: URL(fileURLWithPath: statePath), options: .atomic)

packages/computer-use/cua/native/tests/capabilities_e2e.py

@@ -78,7 +78,8 @@ def main():
             # The Host consent barrier is separately exercised by provider-lifecycle.spec.ts.
             app = call("launchApp", owner)
             pid = app["pid"]
-            state = wait_for_state(state_file, lambda value: value["pid"] == pid)
+            call("activate", owner)
+            state = wait_for_state(state_file, lambda value: value["pid"] == pid and value["active"] and value["keyWindow"])
             assert call("findApps", {"query": str(application), "maxAppMatches": 20})[0]["pid"] == pid
             observed = call("getState", {**owner, "maxElements": 2000, "maxDepth": 25})
             field = next(item for item in observed["elements"] if item["role"] == "AXTextArea")
@@ -120,8 +121,10 @@ def main():
                 pid = json.loads(state_file.read_text())["pid"]
             if args.artifacts:
                 args.artifacts.mkdir(parents=True, exist_ok=True)
+                if state_file.exists():
+                    (args.artifacts / "state-final.json").write_bytes(state_file.read_bytes())
                 (args.artifacts / "transcript.json").write_text(json.dumps(transcript, ensure_ascii=False, indent=2))
-            run_cleanup(*([helper.close] if helper else []), *([lambda: stop_application(pid)] if pid else []))
+            run_cleanup(*([helper.stop] if helper else []), *([lambda: stop_application(pid)] if pid else []))
 
 
 if __name__ == "__main__":