fix(streaming): detect text-only stream drops with no finish_reason (#32086)

When a streaming response ends cleanly (HTTP 200) with no finish_reason
after delivering text but no tool calls, the chunk collector silently
stamps finish_reason='stop' and the conversation loop presents truncated
text as a complete response.

Three stream-drop paths now exist after chunk collection:

1. Zero-chunk → EmptyStreamError, retried (existing)
2. Tool-call in progress, no finish_reason → partial-stream-stub (existing)
3. Text-only, no finish_reason → partial-stream-stub (NEW — this fix)

Path 3 routes through the same PARTIAL_STREAM_STUB_ID + FINISH_REASON_LENGTH
machinery as path 2. The conversation loop shows 'Stream interrupted —
requesting continuation' and injects a continue prompt, giving the model
a chance to resume where the stream dropped.

Observed with DeepSeek provider where CloudFront drops SSE streams
mid-response after delivering partial text.

agent/chat_completion_helpers.py

@@ -2998,6 +2998,40 @@ def interruptible_streaming_api_call(agent, api_kwargs: dict, *, on_first_delta=
                 _dropped_tool_names=_dropped_names or None,
             )
 
+        # Text-only stream drop: the upstream closed the connection (or the
+        # SSE stream simply ended) with no finish_reason after delivering
+        # text content but no tool calls.  Without this guard the partial
+        # text is silently stamped finish_reason="stop" and the turn ends as
+        # if complete — the model's intended next step is lost (#32086).
+        _text_only_dropped_no_finish = (
+            finish_reason is None
+            and content_parts
+            and not tool_calls_acc
+        )
+        if _text_only_dropped_no_finish:
+            logger.warning(
+                "Stream ended with no finish_reason after delivering text "
+                "with no tool calls; treating as a mid-stream drop."
+            )
+            full_reasoning = "".join(reasoning_parts) or None
+            mock_message = SimpleNamespace(
+                role=role,
+                content=full_content,
+                tool_calls=None,
+                reasoning_content=full_reasoning,
+            )
+            mock_choice = SimpleNamespace(
+                index=0,
+                message=mock_message,
+                finish_reason=FINISH_REASON_LENGTH,
+            )
+            return SimpleNamespace(
+                id=PARTIAL_STREAM_STUB_ID,
+                model=model_name,
+                choices=[mock_choice],
+                usage=usage_obj,
+            )
+
         effective_finish_reason = finish_reason or "stop"
         if has_truncated_tool_args:
             effective_finish_reason = "length"

tests/run_agent/test_partial_stream_finish_reason.py

@@ -230,6 +230,45 @@ class TestCleanStreamEndMidToolCall:
         assert response.id.startswith("stream-")
         assert response.choices[0].finish_reason == FINISH_REASON_LENGTH
 
+    @patch("run_agent.AIAgent._create_request_openai_client")
+    @patch("run_agent.AIAgent._close_request_openai_client")
+    def test_no_finish_reason_text_only_routes_to_stub(
+        self, _mock_close, mock_create, monkeypatch,
+    ):
+        """A clean stream-end with no finish_reason after text-only
+        delivery must route through the partial-stream-stub path so the
+        conversation loop continues instead of silently accepting
+        truncated text as a complete response (#32086)."""
+
+        def _clean_ending_stream():
+            yield _make_stream_chunk(content="Let me compare the ")
+            yield _make_stream_chunk(content="vision configs:")
+            # falls off the end — clean close, no terminator
+
+        mock_client = MagicMock()
+        mock_client.chat.completions.create.side_effect = (
+            lambda *a, **kw: _clean_ending_stream()
+        )
+        mock_create.return_value = mock_client
+
+        agent = _make_agent()
+        agent._fire_stream_delta = lambda text: None
+
+        response = agent._interruptible_streaming_api_call({})
+
+        assert response.id == PARTIAL_STREAM_STUB_ID, (
+            "A clean stream-end with no finish_reason after text-only "
+            "delivery must be tagged as a partial-stream stub, not "
+            "silently accepted as complete (#32086)."
+        )
+        assert response.choices[0].finish_reason == FINISH_REASON_LENGTH
+        assert response.choices[0].message.content == "Let me compare the vision configs:"
+        assert response.choices[0].message.tool_calls is None
+        assert getattr(response, "_dropped_tool_names", None) is None, (
+            "Text-only drops must not carry dropped tool names — there "
+            "were no tool calls in flight."
+        )
+
 
 # ── Length-continuation prompt branching ──────────────────────────────────