diff --git a/nemoguardrails/colang/v2_x/library/guardrails.co b/nemoguardrails/colang/v2_x/library/guardrails.co
index 9f6e1de009..bb9640463c 100644
--- a/nemoguardrails/colang/v2_x/library/guardrails.co
+++ b/nemoguardrails/colang/v2_x/library/guardrails.co
@@ -93,7 +93,11 @@ flow run output rails $output_text
   $output_rails_exist = await CheckFlowDefinedAction(flow_id="output rails")
   if $output_rails_exist
     log "found output rails, checking..."
-    await output rails $output_text
-    log "output rails passed"
-
-  $output_rails_in_progress = False
+    when output rails $output_text
+      log "output rails passed"
+      $output_rails_in_progress = False
+    else
+      $output_rails_in_progress = False
+      abort
+  else
+    $output_rails_in_progress = False
diff --git a/tests/v2_x/test_output_rails_abort_multiturn.py b/tests/v2_x/test_output_rails_abort_multiturn.py
new file mode 100644
index 0000000000..1e5a06cafb
--- /dev/null
+++ b/tests/v2_x/test_output_rails_abort_multiturn.py
@@ -0,0 +1,130 @@
+# SPDX-FileCopyrightText: Copyright (c) 2023-2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Regression for #2067: output rails must keep firing after an abort."""
+
+from nemoguardrails import LLMRails, RailsConfig
+from nemoguardrails.utils import new_event_dict, new_uuid
+
+colang_content = """
+import core
+import guardrails
+
+flow output rails $output_text
+  if $output_text == "BLOCKME"
+    bot say "Rejected."
+    abort
+
+@active
+flow handle user messages
+  while True
+    user said something
+    bot say "BLOCKME"
+
+flow main
+  activate handle user messages
+  wait indefinitely
+"""
+
+yaml_content = """
+colang_version: "2.x"
+models: []
+"""
+
+
+def _setup_rails() -> LLMRails:
+    config = RailsConfig.from_content(colang_content, yaml_content)
+    app = LLMRails(config)
+    app.runtime.disable_async_execution = True
+    return app
+
+
+def _finish_bot_actions(app: LLMRails, state, events: list):
+    for event in events:
+        if event["type"] == "StartUtteranceBotAction":
+            state = app.process_events(
+                [
+                    new_event_dict(
+                        "UtteranceBotActionFinished",
+                        action_uid=event["action_uid"],
+                        is_success=True,
+                        final_script=event["script"],
+                    )
+                ],
+                state,
+            )[1]
+    return state
+
+
+def _user_said(text: str) -> list:
+    uid = new_uuid()
+    return [
+        new_event_dict("UtteranceUserActionStarted", action_uid=uid),
+        new_event_dict(
+            "UtteranceUserActionFinished",
+            final_transcript=text,
+            action_uid=uid,
+            is_success=True,
+        ),
+    ]
+
+
+def _bot_scripts(events: list) -> list[str]:
+    return [event["script"] for event in events if event["type"] == "StartUtteranceBotAction"]
+
+
+def test_output_rails_fire_on_every_turn_after_abort():
+    """After an output-rail abort, later turns must still run output rails."""
+    app = _setup_rails()
+    state = None
+
+    for turn in ("hello", "hi again", "one more"):
+        out, state = app.process_events(_user_said(turn), state)
+        state = _finish_bot_actions(app, state, out)
+        scripts = _bot_scripts(out)
+        assert scripts == ["Rejected."], f"turn {turn!r}: expected output rail to reject BLOCKME, got {scripts!r}"
+
+
+def test_output_rails_pass_through_still_clears_flag():
+    """Successful output rails must clear the flag so later bot says still work."""
+    colang = """
+import core
+import guardrails
+
+flow output rails $output_text
+  global $bot_message
+  $bot_message = "ok:{$output_text}"
+
+@active
+flow handle user messages
+  while True
+    user said something
+    bot say "passthrough"
+
+flow main
+  activate handle user messages
+  wait indefinitely
+"""
+    config = RailsConfig.from_content(colang, yaml_content)
+    app = LLMRails(config)
+    app.runtime.disable_async_execution = True
+    state = None
+
+    for turn in ("alpha", "beta"):
+        out, state = app.process_events(_user_said(turn), state)
+        state = _finish_bot_actions(app, state, out)
+        assert _bot_scripts(out) == ["ok:passthrough"], (
+            f"turn {turn!r}: expected transformed output, got {_bot_scripts(out)!r}"
+        )