diff --git a/tests/unittests/a2a/utils/test_agent_card_builder.py b/tests/unittests/a2a/utils/test_agent_card_builder.py
index c979ad53..dc4dddd4 100644
--- a/tests/unittests/a2a/utils/test_agent_card_builder.py
+++ b/tests/unittests/a2a/utils/test_agent_card_builder.py
@@ -285,11 +285,7 @@ class TestAgentCardBuilder:
     assert any("writer" in sid for sid in skill_ids)  # child node skill
 
   async def test_build_succeeds_for_workflow_with_output_schema_node(self):
-    """AgentCardBuilder.build succeeds for a Workflow whose LlmAgent has output_schema.
-
-    Mirrors the exact repro from
-    https://github.com/google/adk-python/issues/5487.
-    """
+    """AgentCardBuilder.build succeeds for a Workflow whose LlmAgent has output_schema."""
 
     class _Out(BaseModel):
       text: str
diff --git a/tests/unittests/artifacts/test_artifact_service.py b/tests/unittests/artifacts/test_artifact_service.py
index e637ccba..a9802e66 100644
--- a/tests/unittests/artifacts/test_artifact_service.py
+++ b/tests/unittests/artifacts/test_artifact_service.py
@@ -1145,10 +1145,7 @@ async def test_gcs_load_artifact_file_data_fallback_compatibility() -> None:
 async def test_save_artifact_with_camel_case_dict(
     service_type, artifact_service_factory
 ):
-  """Artifact services accept camelCase dicts (Agentspace format).
-
-  Regression test for https://github.com/google/adk-python/issues/2886
-  """
+  """Artifact services accept camelCase dicts (Agentspace format)."""
   artifact_service = artifact_service_factory(service_type)
   app_name = "app0"
   user_id = "user0"
diff --git a/tests/unittests/flows/llm_flows/test_functions_thread_pool.py b/tests/unittests/flows/llm_flows/test_functions_thread_pool.py
index 33a9af46..50978cb2 100644
--- a/tests/unittests/flows/llm_flows/test_functions_thread_pool.py
+++ b/tests/unittests/flows/llm_flows/test_functions_thread_pool.py
@@ -342,7 +342,6 @@ class TestCallToolInThreadPool:
   ):
     """FunctionTools returning None or other falsy values must execute exactly once.
 
-    Regression test for https://github.com/google/adk-python/issues/5284.
     Previously, a None return was mistaken for the internal sentinel used to
     signal 'non-FunctionTool, fall back to run_async', causing a second
     invocation. The fix uses an identity-based sentinel so that None and other
diff --git a/tests/unittests/flows/llm_flows/test_llm_callback_span_consistency.py b/tests/unittests/flows/llm_flows/test_llm_callback_span_consistency.py
index cb7a0e54..bc19c458 100644
--- a/tests/unittests/flows/llm_flows/test_llm_callback_span_consistency.py
+++ b/tests/unittests/flows/llm_flows/test_llm_callback_span_consistency.py
@@ -12,10 +12,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-"""Tests that before/after/error model callbacks all observe the same call_llm span.
-
-Regression tests for https://github.com/google/adk-python/issues/4851.
-"""
+"""Tests that before/after/error model callbacks all observe the same call_llm span."""
 
 from typing import AsyncGenerator
 from typing import Optional
diff --git a/tests/unittests/plugins/test_bigquery_agent_analytics_plugin.py b/tests/unittests/plugins/test_bigquery_agent_analytics_plugin.py
index 7cc8e360..fc38e9d5 100644
--- a/tests/unittests/plugins/test_bigquery_agent_analytics_plugin.py
+++ b/tests/unittests/plugins/test_bigquery_agent_analytics_plugin.py
@@ -5301,7 +5301,7 @@ class TestHITLTracing:
 
 
 class TestSpanHierarchyIsolation:
-  """Regression tests for https://github.com/google/adk-python/issues/4561.
+  """Regression tests for span hierarchy isolation.
 
   ``push_span()`` must NOT attach its span to the ambient OTel context.
   If it does, any subsequent ``tracer.start_as_current_span()`` in the
@@ -6119,8 +6119,6 @@ class TestAnalyticsViews:
 class TestTraceIdContinuity:
   """Tests for trace_id continuity across all events in an invocation.
 
-  Regression tests for https://github.com/google/adk-python/issues/4645.
-
   When there is no ambient OTel span (e.g. Agent Engine, custom runners),
   early events (USER_MESSAGE_RECEIVED, INVOCATION_STARTING) used to fall
   back to ``invocation_id`` while AGENT_STARTING got a new OTel hex
diff --git a/tests/unittests/sessions/test_session_service.py b/tests/unittests/sessions/test_session_service.py
index b054451c..157e4fb2 100644
--- a/tests/unittests/sessions/test_session_service.py
+++ b/tests/unittests/sessions/test_session_service.py
@@ -1468,8 +1468,7 @@ async def test_service_recovers_after_multiple_failures():
 @pytest.mark.asyncio
 async def test_concurrent_prepare_tables_no_race_condition():
   """Verifies that concurrent calls to prepare_tables wait for table creation.
-  Reproduces the race condition from
-  https://github.com/google/adk-python/issues/4445: when concurrent requests
+  Reproduces the race condition where concurrent requests
   arrive at startup, prepare_tables must not return before tables exist.
   Previously, the early-return guard checked _db_schema_version (set during
   schema detection) instead of _tables_created, so a second request could
@@ -1602,7 +1601,7 @@ async def test_get_or_create_state_creates_new_row():
 async def test_get_or_create_state_handles_race_condition():
   """_get_or_create_state recovers when a concurrent INSERT wins the race.
 
-  Simulates the race from https://github.com/google/adk-python/issues/4954:
+  Simulates the race:
   the initial SELECT returns None (another caller hasn't committed yet), but
   by the time we INSERT, the other caller has committed — so the INSERT fails
   with IntegrityError and we fall back to re-fetching.
diff --git a/tests/unittests/tools/openapi_tool/openapi_spec_parser/test_rest_api_tool.py b/tests/unittests/tools/openapi_tool/openapi_spec_parser/test_rest_api_tool.py
index 412d16f6..ad72915e 100644
--- a/tests/unittests/tools/openapi_tool/openapi_spec_parser/test_rest_api_tool.py
+++ b/tests/unittests/tools/openapi_tool/openapi_spec_parser/test_rest_api_tool.py
@@ -1130,7 +1130,6 @@ class TestRestApiTool:
     httpx defaults to a 5-second timeout, which is too short for many
     real-world API calls. Verify that we explicitly disable the timeout
     to match the previous requests-library behavior (no timeout).
-    Regression test for https://github.com/google/adk-python/issues/4431.
     """
     mock_response = mock.create_autospec(requests.Response, instance=True)
     mock_response.json.return_value = {"result": "success"}
@@ -1455,7 +1454,6 @@ class TestRestApiTool:
     in the OpenAPI path (e.g. '...execute?triggerId=api_trigger/Name#action').
     These must be moved into the explicit query_params dict so httpx does not
     strip them when it replaces the URL query string with the `params` arg.
-    Regression test for https://github.com/google/adk-python/issues/4555.
     """
     integration_path = (
         "/v2/projects/my-proj/locations/us-central1"
diff --git a/tests/unittests/utils/test_streaming_utils.py b/tests/unittests/utils/test_streaming_utils.py
index 61ce590d..eb92c6d6 100644
--- a/tests/unittests/utils/test_streaming_utils.py
+++ b/tests/unittests/utils/test_streaming_utils.py
@@ -485,10 +485,7 @@ class TestStreamingResponseAggregator:
 
 
 class TestFunctionCallIdGeneration:
-  """Tests for function call ID generation in streaming mode.
-
-  Regression tests for https://github.com/google/adk-python/issues/4609.
-  """
+  """Tests for function call ID generation in streaming mode."""
 
   @pytest.mark.asyncio
   async def test_non_streaming_fc_generates_id_when_empty(self):