diff --git a/llama_stack/distribution/inspect.py b/llama_stack/distribution/inspect.py
index 5822070ad..d6a598982 100644
--- a/llama_stack/distribution/inspect.py
+++ b/llama_stack/distribution/inspect.py
@@ -54,6 +54,7 @@ class DistributionInspectImpl(Inspect):
                             provider_types=[],  # These APIs don't have "real" providers - they're internal to the stack
                         )
                         for e in endpoints
+                        if e.methods is not None
                     ]
                 )
             else:
@@ -67,6 +68,7 @@ class DistributionInspectImpl(Inspect):
                                 provider_types=[p.provider_type for p in providers],
                             )
                             for e in endpoints
+                            if e.methods is not None
                         ]
                     )
 
diff --git a/llama_stack/distribution/library_client.py b/llama_stack/distribution/library_client.py
index 5dc0078d4..6503c13b2 100644
--- a/llama_stack/distribution/library_client.py
+++ b/llama_stack/distribution/library_client.py
@@ -445,8 +445,9 @@ class AsyncLlamaStackAsLibraryClient(AsyncLlamaStackClient):
         # we use asynchronous impl always internally and channel all requests to AsyncLlamaStackClient
         # however, the top-level caller may be a SyncAPIClient -- so its stream_cls might be a Stream (SyncStream)
         # so we need to convert it to AsyncStream
+        # mypy can't track runtime variables inside the [...] of a generic, so ignore that check
         args = get_args(stream_cls)
-        stream_cls = AsyncStream[args[0]]
+        stream_cls = AsyncStream[args[0]]  # type: ignore[valid-type]
         response = AsyncAPIResponse(
             raw=mock_response,
             client=self,