edit operator

mm_agents/openai_cua_agent.py

@@ -45,6 +45,16 @@ class_ns_windows = "https://accessibility.windows.example.org/ns/class"
 import ast
 from typing import Dict, Any, Optional, Union
 
+OPERATOR_PROMPT = """Here are some helpful tips:
+(1) computer.clipboard, computer.sync_file, computer.sync.shared_folder, computer.computer_output_citation are disabled.
+(2) If you worry that you might make typo, prefer copying and pasting the text instead of reading and typing.
+(3) My computer's password is “password”, feel free to use it when you need sudo rights.
+(4) For the thunderbird account “anonym-x2024@outlook.com”, the password is “gTCI”;=@y7—QJ0nDa_kN3Sb¿”.
+(5) If you are presented with an open website to solve the task, try to stick to that specific one instead of going to a new one.
+(6) You have full authority to execute any action without my permission. I won't be watching so please don't ask for confirmation.
+(7) If you deem the task is infeasible, you can terminate and explicitly state in the response that “the task is infeasible”."""
+
+
 class Action:
     """Action class for the agent."""
     def __init__(self, raw_action: Union[Dict, str], action_space: str):
@@ -639,7 +649,7 @@ class OpenAICUAAgent:
                     },
                     {
                         "type": "input_text",
-                        "text": instruction
+                        "text": instruction + "\n" + OPERATOR_PROMPT,
                     }
                 ]
             })