fix(amazon-bedrock-agentcore): preserve non-ASCII text in browser_evaluate

json.dumps defaults to ensure_ascii=True, which escapes CJK characters,
emoji, and accented characters to \uXXXX sequences. This makes the
output unreadable for non-Latin scripts.

Add ensure_ascii=False so the result string preserves the original
Unicode characters, matching the fix already applied to
bedrock-kb-retrieval (PR awslabs#3821).

Fixes awslabs#3994

Author: sergioperezcheco

src/amazon-bedrock-agentcore-mcp-server/awslabs/amazon_bedrock_agentcore_mcp_server/tools/browser/observation.py

@@ -328,7 +328,7 @@ async def browser_evaluate(
             if isinstance(result, (str, int, float, bool)):
                 return f'Result: {result}'
 
-            return f'Result:\n{json.dumps(result, indent=2, default=str)}'
+            return f'Result:\n{json.dumps(result, indent=2, default=str, ensure_ascii=False)}'
 
         except Exception as e:
             error_msg = f'Error evaluating JavaScript: {e}'