[nrfconnect] Fix all-clusters-app commissioning (#17399)

Common ZAP for all-clusters-app has another endpoint 0xFFFE
with NetworkCommissioning cluster, but that is not used by
Thread-based platforms and causes failures when wildcard
reads are used.

examples/all-clusters-app/nrfconnect/main/AppTask.cpp

@@ -39,6 +39,10 @@
 #include <logging/log.h>
 #include <zephyr.h>
 
+using namespace ::chip;
+using namespace ::chip::Credentials;
+using namespace ::chip::DeviceLayer;
+
 #define FACTORY_RESET_TRIGGER_TIMEOUT 3000
 #define FACTORY_RESET_CANCEL_WINDOW_TIMEOUT 3000
 #define APP_EVENT_QUEUE_SIZE 10
@@ -52,6 +56,8 @@ static LEDWidget sStatusLED;
 static UnusedLedsWrapper<3> sUnusedLeds{ { DK_LED2, DK_LED3, DK_LED4 } };
 static k_timer sFunctionTimer;
 
+constexpr EndpointId kNetworkCommissioningEndpointSecondary = 0xFFFE;
+
 namespace LedConsts {
 constexpr uint32_t kBlinkRate_ms{ 500 };
 namespace StatusLed {
@@ -67,10 +73,6 @@ constexpr uint32_t kOff_ms{ 950 };
 } // namespace StatusLed
 } // namespace LedConsts
 
-using namespace ::chip;
-using namespace ::chip::Credentials;
-using namespace ::chip::DeviceLayer;
-
 CHIP_ERROR AppTask::Init()
 {
     // Initialize CHIP stack
@@ -137,6 +139,8 @@ CHIP_ERROR AppTask::Init()
 #if CONFIG_CHIP_OTA_REQUESTOR
     InitBasicOTARequestor();
 #endif
+    // We only have network commissioning on endpoint 0.
+    emberAfEndpointEnableDisable(kNetworkCommissioningEndpointSecondary, false);
     ConfigurationMgr().LogDeviceConfig();
     PrintOnboardingCodes(chip::RendezvousInformationFlag(chip::RendezvousInformationFlag::kBLE));