fix(agent): clear stale intermediate acknowledgments

Treat intent-ack continuation text as non-final so last-turn exhaustion requests a real summary instead of surfacing a premature promise. Keep iteration-limit fallback text free of the abnormal-fragment explainer.

agent/conversation_loop.py

@@ -5104,6 +5104,10 @@ def run_conversation(
                     }
                     messages.append(continue_msg)
                     agent._session_messages = messages
+                    # An acknowledgment is explicitly non-final. Do not let its
+                    # text suppress iteration-limit summarization if this
+                    # continuation consumes the remaining budget.
+                    final_response = None
                     continue
 
                 codex_ack_continuations = 0

agent/turn_finalizer.py

@@ -320,7 +320,7 @@ def finalize_turn(
                 # truncated partial (the "The" case from #34452).
                 _is_partial_fragment = (
                     not _is_empty_terminal
-                    and not continuation_budget_exhausted
+                    and not iteration_limit_fallback
                     and not str(_turn_exit_reason).startswith("text_response")
                     and len(_stripped) <= 24
                     and _stripped[-1:] not in {".", "!", "?", "。", "！", "？", "`", ")"}

tests/run_agent/test_verification_continuation_budget.py

@@ -102,3 +102,22 @@ def test_pre_verify_preserves_composed_report_at_budget_limit(agent, monkeypatch
     _assert_pending_response_survives(agent, result)
     assert result["messages"][1]["_pre_verify_synthetic"] is True
     assert result["messages"][2]["_pre_verify_synthetic"] is True
+
+
+def test_intermediate_ack_uses_summary_instead_of_premature_text(agent, monkeypatch):
+    agent.valid_tool_names = ["web_search"]
+    agent._intent_ack_continuation = True
+    agent._looks_like_codex_intermediate_ack = MagicMock(return_value=True)
+    agent._interruptible_api_call = lambda _kwargs: _response("I'll inspect the files now")
+    agent._handle_max_iterations = MagicMock(return_value="verified summary")
+    monkeypatch.setenv("HERMES_VERIFY_ON_STOP", "0")
+
+    with (
+        patch("hermes_cli.plugins.has_hook", return_value=False),
+        patch("hermes_cli.plugins.invoke_hook", return_value=[]),
+    ):
+        result = agent.run_conversation("inspect /tmp/project")
+
+    assert result["final_response"] == "verified summary"
+    assert result["turn_exit_reason"] == "max_iterations_reached(1/1)"
+    agent._handle_max_iterations.assert_called_once()