chore: tighten runner fallback semantics and test naming

Agent-Logs-Url: https://github.com/bg-playground/BGSTM/sessions/5c293b39-4619-4432-b5ac-5f928ca1b000

Co-authored-by: bg-playground <259109604+bg-playground@users.noreply.github.com>

Author: Copilot

backend/app/api/external_results.py

@@ -73,7 +73,9 @@ async def create_external_session(
     Returns the existing session if an identical session was created within the
     last 60 seconds (idempotency window).
     """
-    normalized_payload = payload.model_copy(update={"runner": payload.runner or _DEFAULT_RUNNER})
+    normalized_payload = payload.model_copy(
+        update={"runner": payload.runner if payload.runner is not None else _DEFAULT_RUNNER}
+    )
     session = await create_session(db, payload=normalized_payload, runner_token_id=token.id)
     await write_audit(
         db,

backend/tests/integration/test_external_results_session.py

@@ -151,7 +151,7 @@ def test_create_finish_fetch(self, db_session, write_token):
             assert data3["status"] == "passed"
             assert data3["id"] == session_id
 
-    def test_create_without_runner_uses_default_runner(self, db_session, write_token):
+    def test_create_without_runner_defaults_to_bgstm_playwright_core(self, db_session, write_token):
         _model, plaintext = write_token
         headers = _auth_header(plaintext)
         payload = dict(_SESSION_PAYLOAD)