diff --git a/evalharness/agent/envs/tau2_official.py b/evalharness/agent/envs/tau2_official.py
index 2f4413c..015352d 100644
--- a/evalharness/agent/envs/tau2_official.py
+++ b/evalharness/agent/envs/tau2_official.py
@@ -119,6 +119,14 @@ def _patch_tau2_generate(adapter, user_adapter=None, gen_kwargs=None) -> None:
 
         import json as _json
 
+        # strip thinking leakage: hybrid backends inline reasoning in
+        # content as '...scenario reasoning...</think>visible reply' -- the
+        # USER simulator's leak hands the agent the scenario's secret
+        # instructions (task goal, what to disclose), inflating rewards
+        text = out.text or ''
+        if '</think>' in text:
+            text = text.rsplit('</think>', 1)[-1].strip()
+
         calls = []
         for c in out.tool_calls:
             try:
@@ -127,7 +135,7 @@ def _patch_tau2_generate(adapter, user_adapter=None, gen_kwargs=None) -> None:
                 args = {'raw': c.arguments}
             calls.append(TauToolCall(id=c.id or f'call_{c.name}', name=c.name,
                                      arguments=args or {}))
-        return AssistantMessage(role='assistant', content=out.text or '',
+        return AssistantMessage(role='assistant', content=text,
                                 tool_calls=calls or None, cost=None,
                                 usage=None, raw_data=None)