test: trim Responses blank-carrier coverage to two invariants

Keep the fixture-driven test on the real failing turn shape from #103483
(no invented carrier, every reasoning item followed by its function_call,
call/output pairs intact) and the control that a lone reasoning item still
gets a non-empty follower. Drop the synthetic parametrized round: the fixture
already covers that shape once per tool round, and the repo caps a fix at two
invariant tests.

Part of #103483

tests/agent/test_responses_empty_tool_replay.py

@@ -1,31 +1,18 @@
 """Responses replay must not invent a blank assistant turn before a tool call.
 
-Regression investigation for #103483: synthetic Muse tool-loop A/B reproduction.
-Also covers #75202: the lone-reasoning follower must be non-empty for strict
-Responses-compatible providers that reject "" with 400.
+Regression tests for #103483 (Muse Spark degenerate finals on the Responses
+wire) and #75202 (strict Responses-compatible providers reject an empty
+``content`` string with 400).
 """
 import json
 from pathlib import Path
 
-import pytest
 from agent.codex_responses_adapter import _chat_messages_to_responses_input
 
-@pytest.mark.parametrize('content', ['', None])
-def test_reasoning_tool_round_has_no_invented_assistant_message(content):
-    messages = [{
-        'role': 'assistant', 'content': content,
-        'codex_reasoning_items': [{'type': 'reasoning', 'id': 'rs_test',
-                                  'encrypted_content': 'synthetic-encrypted-fixture', 'summary': []}],
-        'tool_calls': [{'id': 'call_test', 'type': 'function',
-                        'function': {'name': 'record_step', 'arguments': '{"step":1}'}}],
-    }, {'role': 'tool', 'tool_call_id': 'call_test', 'content': '{"ok":true}'}]
-    items = _chat_messages_to_responses_input(messages)
-    assert [item.get('type', item.get('role')) for item in items] == [
-        'reasoning', 'function_call', 'function_call_output']
-    assert items[1]['call_id'] == items[2]['call_id']
-
 
 def test_reasoning_without_tool_keeps_nonempty_following_item():
+    """Control: a lone reasoning item still needs a follower (else
+    ``missing_following_item``), and it must be non-empty for strict providers."""
     items = _chat_messages_to_responses_input([{
         'role': 'assistant', 'content': '',
         'codex_reasoning_items': [{'type': 'reasoning', 'id': 'rs_test',