diff --git a/doc/rules.md b/doc/rules.md
index ec9e4d091..5745ba4de 100644
--- a/doc/rules.md
+++ b/doc/rules.md
@@ -726,10 +726,20 @@ root of your workspace (i.e. `$(SRCROOT)`).
 
 ### Test Filtering
 
-A subset of tests for a given target can be executed via the `--test_filter` parameter:
+`swift_test` supports Bazel's `--test_filter` flag on all platforms (i.e., Apple
+and Linux), which can be used to run only a subset of tests. The expected filter
+format is the same as Xcode's `xctest` tool:
 
-```
-bazel test //:Tests --test_filter=TestModuleName.TestClassName/testMethodName
+*   `ModuleName`: Run only the test classes/methods in module `ModuleName`.
+*   `ModuleName.ClassName`: Run only the test methods in class
+    `ModuleName.ClassName`.
+*   `ModuleName.ClassName/testMethodName`: Run only the method `testMethodName`
+    in class `ModuleName.ClassName`.
+
+Multiple such filters can be separated by commas. For example:
+
+```shell
+bazel test --test_filter=AModule,BModule.SomeTests,BModule.OtherTests/testX //my/package/...
 ```
 
 **ATTRIBUTES**
diff --git a/examples/apple/test_filter/BUILD b/examples/xplatform/test_filter/BUILD
similarity index 76%
rename from examples/apple/test_filter/BUILD
rename to examples/xplatform/test_filter/BUILD
index 9d122bd9f..c251aa05a 100644
--- a/examples/apple/test_filter/BUILD
+++ b/examples/xplatform/test_filter/BUILD
@@ -11,7 +11,6 @@ swift_library(
         "TestHelper.swift",
     ],
     module_name = "test_filter",
-    target_compatible_with = ["@platforms//os:macos"],
 )
 
 # Verify that tests fail as expected without test filtering.
@@ -20,8 +19,7 @@ swift_test(
     env = {
         "EXPECT_FAILURE": "TRUE",
     },
-    module_name = "test_filter",
-    target_compatible_with = ["@platforms//os:macos"],
+    module_name = "test_filter__baseline",
     deps = [":test_filter_lib"],
 )
 
@@ -31,8 +29,7 @@ swift_test(
     env = {
         "TESTBRIDGE_TEST_ONLY": "test_filter.PassTests",
     },
-    module_name = "test_filter",
-    target_compatible_with = ["@platforms//os:macos"],
+    module_name = "test_filter__feature__target_class",
     deps = [":test_filter_lib"],
 )
 
@@ -42,7 +39,6 @@ swift_test(
     env = {
         "TESTBRIDGE_TEST_ONLY": "test_filter.PassFailTests/test_pass",
     },
-    module_name = "test_filter",
-    target_compatible_with = ["@platforms//os:macos"],
+    module_name = "test_filter__feature__target_class_method",
     deps = [":test_filter_lib"],
 )
diff --git a/examples/apple/test_filter/FailTests.swift b/examples/xplatform/test_filter/FailTests.swift
similarity index 100%
rename from examples/apple/test_filter/FailTests.swift
rename to examples/xplatform/test_filter/FailTests.swift
diff --git a/examples/apple/test_filter/PassFailTests.swift b/examples/xplatform/test_filter/PassFailTests.swift
similarity index 100%
rename from examples/apple/test_filter/PassFailTests.swift
rename to examples/xplatform/test_filter/PassFailTests.swift
diff --git a/examples/apple/test_filter/PassTests.swift b/examples/xplatform/test_filter/PassTests.swift
similarity index 100%
rename from examples/apple/test_filter/PassTests.swift
rename to examples/xplatform/test_filter/PassTests.swift
diff --git a/examples/apple/test_filter/TestHelper.swift b/examples/xplatform/test_filter/TestHelper.swift
similarity index 100%
rename from examples/apple/test_filter/TestHelper.swift
rename to examples/xplatform/test_filter/TestHelper.swift
diff --git a/swift/swift_test.bzl b/swift/swift_test.bzl
index ede3a350b..e97b49d29 100644
--- a/swift/swift_test.bzl
+++ b/swift/swift_test.bzl
@@ -709,10 +709,20 @@ root of your workspace (i.e. `$(SRCROOT)`).
 
 ### Test Filtering
 
-A subset of tests for a given target can be executed via the `--test_filter` parameter:
+`swift_test` supports Bazel's `--test_filter` flag on all platforms (i.e., Apple
+and Linux), which can be used to run only a subset of tests. The expected filter
+format is the same as Xcode's `xctest` tool:
 
-```
-bazel test //:Tests --test_filter=TestModuleName.TestClassName/testMethodName
+*   `ModuleName`: Run only the test classes/methods in module `ModuleName`.
+*   `ModuleName.ClassName`: Run only the test methods in class
+    `ModuleName.ClassName`.
+*   `ModuleName.ClassName/testMethodName`: Run only the method `testMethodName`
+    in class `ModuleName.ClassName`.
+
+Multiple such filters can be separated by commas. For example:
+
+```shell
+bazel test --test_filter=AModule,BModule.SomeTests,BModule.OtherTests/testX //my/package/...
 ```
 """,
     executable = True,
diff --git a/tools/test_discoverer/TestPrinter.swift b/tools/test_discoverer/TestPrinter.swift
index 4b9fc3e79..d9653489d 100644
--- a/tools/test_discoverer/TestPrinter.swift
+++ b/tools/test_discoverer/TestPrinter.swift
@@ -144,6 +144,7 @@ struct TestPrinter {
 
     var contents = """
       import BazelTestObservation
+      import Foundation
       import XCTest
 
       @main
@@ -162,11 +163,26 @@ struct TestPrinter {
         """
     }
 
+    // Bazel's `--test_filter` flag is passed to the test binary via the `TESTBRIDGE_TEST_ONLY`
+    // environment variable. Below, we read that value if it is present and pass it to `XCTMain`,
+    // where it behaves the same as the `-XCTest` flag value in Xcode's `xctest` tool.
+    //
+    // Note that `XCTMain` treats `arguments` as an actual command line, meaning that `arguments[0]`
+    // is expected to be the binary's path. So we only do the test filter logic if we've already
+    // been passed no other arguments. This lets the test binary still support other XCTest flags
+    // like `--list-tests` or `--dump-tests-json` if the user wishes to use those directly.
     contents += """
           if let xmlObserver = BazelXMLTestObserver.default {
             XCTestObservationCenter.shared.addTestObserver(xmlObserver)
           }
-          XCTMain(tests)
+
+          var arguments = CommandLine.arguments
+          if arguments.count == 1,
+            let testFilter = ProcessInfo.processInfo.environment["TESTBRIDGE_TEST_ONLY"]
+          {
+            arguments.append(testFilter)
+          }
+          XCTMain(tests, arguments: arguments)
         }
       }
 
diff --git a/tools/xctest_runner/xctest_runner.sh.template b/tools/xctest_runner/xctest_runner.sh.template
index 746c2d0f6..a198ae982 100644
--- a/tools/xctest_runner/xctest_runner.sh.template
+++ b/tools/xctest_runner/xctest_runner.sh.template
@@ -17,7 +17,7 @@
 set -euo pipefail
 
 # This test script is used as the executable output of a `swift_test` rule when
-# building macOS targets (unless the "swift.bundled_tests" feature is disabled)
+# building macOS targets (unless the target has `discover_tests = False`)
 # because the test binary is an `MH_BUNDLE` that needs to be loaded dynamically
 # and runtime reflection is used to locate the test methods.