diff --git a/backends/arm/test/conftest.py b/backends/arm/test/conftest.py
index 351d6de7a09..e949247068f 100644
--- a/backends/arm/test/conftest.py
+++ b/backends/arm/test/conftest.py
@@ -29,7 +29,30 @@ def pytest_configure(config):
 
 
 def pytest_collection_modifyitems(config, items):
-    pass
+    from executorch.backends.arm.test import common
+    from executorch.backends.arm.test.runner_utils import model_converter_installed
+
+    skip_marker = getattr(
+        common.SkipIfNoModelConverter, "mark", common.SkipIfNoModelConverter
+    )
+
+    if not model_converter_installed():
+        deselected = []
+        remaining = []
+        for item in items:
+            for marker in item.iter_markers("skipif"):
+                if (
+                    marker.name == skip_marker.name
+                    and marker.args == skip_marker.args
+                    and marker.kwargs == skip_marker.kwargs
+                ):
+                    deselected.append(item)
+                    break
+            else:
+                remaining.append(item)
+        if deselected:
+            config.hook.pytest_deselected(items=deselected)
+            items[:] = remaining
 
 
 def pytest_addoption(parser):
diff --git a/backends/vulkan/test/test_vulkan_delegate.py b/backends/vulkan/test/test_vulkan_delegate.py
index 7c9f31b720c..ac719ad3ade 100644
--- a/backends/vulkan/test/test_vulkan_delegate.py
+++ b/backends/vulkan/test/test_vulkan_delegate.py
@@ -7,6 +7,7 @@
 # pyre-unsafe
 
 import ctypes
+import os
 import unittest
 from typing import Tuple
 
@@ -38,10 +39,35 @@
 
 try:
     ctypes.CDLL("libvulkan.so.1")
-except:
+except OSError:
     pass
 
 
+def _is_using_swiftshader() -> bool:
+    try:
+        if os.environ.get("ETVK_USING_SWIFTSHADER", "0") in ("1", "True"):
+            return True
+        vk_icd = os.environ.get("VK_ICD_FILENAMES", "")
+        if "swiftshader" in vk_icd.lower():
+            return True
+        RTLD_NOLOAD = 4
+        for lib_name in ("libvk_swiftshader.so", "libvk_swiftshader_fbcode.so"):
+            try:
+                ctypes.CDLL(lib_name, mode=RTLD_NOLOAD)
+                return True
+            except Exception:
+                continue
+        return False
+    except Exception:
+        return True
+
+
+skip_if_swiftshader = unittest.skipIf(
+    _is_using_swiftshader(),
+    "Not compatible with swiftshader",
+)
+
+
 def lower_module(
     model: torch.nn.Module, sample_inputs: Tuple[torch.Tensor], dynamic_shapes=None
 ) -> EdgeProgramManager:
@@ -590,6 +616,7 @@ def forward(self, x):
 
         self.lower_unary_module_and_test_output(SqrtModule())
 
+    @skip_if_swiftshader
     def test_vulkan_backend_hardshrink(self):
         class HardshrinkModule(torch.nn.Module):
             def __init__(self):
@@ -1028,7 +1055,7 @@ def forward(self, x):
             sample_inputs,
         )
 
-    @unittest.skip("layer norm compute shader not working with swiftshader")
+    @skip_if_swiftshader
     def test_vulkan_backend_native_layer_norm(self):
         class NativeLayerNormModule(torch.nn.Module):
             def __init__(self):
@@ -1459,9 +1486,7 @@ def forward(self, x):
             sample_inputs,
         )
 
-    @unittest.skip(
-        "Softmax shader with shared memory does not work with swiftshader due to potential swiftshader bug"
-    )
+    @skip_if_swiftshader
     def test_vulkan_backend_softmax(self):
         class SoftmaxModule(torch.nn.Module):
             def __init__(self):
@@ -1480,9 +1505,7 @@ def forward(self, x):
             sample_inputs,
         )
 
-    @unittest.skip(
-        "Softmax shader with shared memory does not work with swiftshader due to potential swiftshader bug"
-    )
+    @skip_if_swiftshader
     def test_vulkan_backend_logsoftmax(self):
         class LogSoftmaxModule(torch.nn.Module):
             def __init__(self):
@@ -2364,7 +2387,7 @@ def apply_quantization(self):
             quantized_linear_module_gemm, sample_inputs_gemm, atol=1e-2, rtol=1e-2
         )
 
-    @unittest.skip("Cannot run on swiftshader due to no integer dot product support")
+    @skip_if_swiftshader
     def test_vulkan_backend_xnnpack_pt2e_quantized_linear_sequence(self):
         """
         Test a sequence of linear layers quantized with XNNPACK quantization config.
@@ -2439,7 +2462,7 @@ def forward(self, x):
             rtol=1e-1,
         )
 
-    @unittest.skip("Cannot run on swiftshader due to no integer dot product support")
+    @skip_if_swiftshader
     def test_vulkan_backend_xnnpack_pt2e_quantized_conv_sequence(self):
         """
         Test a sequence of convolution layers quantized with PT2E quantization.
@@ -2530,7 +2553,7 @@ def forward(self, x):
             rtol=1e-1,
         )
 
-    @unittest.skip("Cannot run on swiftshader due to no integer dot product support")
+    @skip_if_swiftshader
     def test_vulkan_backend_xnnpack_pt2e_quantized_conv_sequence_all_reduced(self):
         """
         Test a sequence of convolution layers quantized with PT2E quantization.
@@ -2610,7 +2633,7 @@ def forward(self, x):
             rtol=1e-1,
         )
 
-    @unittest.skip("Cannot run on swiftshader due to no 8-bit int support")
+    @skip_if_swiftshader
     def test_vulkan_backend_torchao_8da4w_quantized_linear(self):
         """
         Test TorchAO 8da4w quantization (int8 dynamic activation + int4 weight) with Vulkan backend.
diff --git a/exir/backend/test/test_partitioner.py b/exir/backend/test/test_partitioner.py
index dedcfe52966..682554edee7 100644
--- a/exir/backend/test/test_partitioner.py
+++ b/exir/backend/test/test_partitioner.py
@@ -106,7 +106,7 @@ def partition(
 
         with self.assertRaisesRegex(
             AttributeError,
-            "can't set attribute 'spec'",
+            "can't set attribute 'spec'|has no setter",
         ):
             my_partitioner.spec = {"new_key": "new_value"}