Command execute_command_chain returned: Unknown command 'execute_command_chain'. Please refer to the 'COMMANDS' list for available commands and only respond in the specified JSON format.

[GhostfireZA]

Duplicates

• I have searched the existing issues

Steps to reproduce 🕹

Just ask it to clone a git repository

Goal1: Clone, review, and clean the code: Clone the target project from my GitHub account at [GitHub here] and analyze the source code and remove any unused or unnecessary code elements.

Current behavior 😯

PLAN:
-  Clone the GitHub repository using the 'git clone' command.
-  Verify that the clone was successful using the 'ls' command.

CRITICISM: We should remember to navigate to the desired directory before executing any the commands.

NEXT ACTION:  COMMAND = execute_command_chain ARGUMENTS = {'commands': [{'name': 'change_directory', 'args': {'directory': '~/'}}, {'name': 'make_directory', 'args': {'directory': 'projectX'}}, {'name': 'change_directory', 'args': {'directory': 'projectX'}}, {'name': 'git_clone', 'args': {'repository': '[The repo you specified]'}}, {'name': 'list_directory', 'args': {'directory': '.'}}]}

SYSTEM:  Command execute_command_chain returned: Unknown command 'execute_command_chain'. Please refer to the 'COMMANDS' list for available commands and only respond in the specified JSON format.

Expected behavior 🤔

Well, it should not break

Your prompt 📝

Name:  Maker
Role:  a autonomous AI that acts as a Code Reviewer. Your main goals are to clone a specific GitHub repository.
Goals: ['Clone the target project from my GitHub account at https://github.com/Torantulino/Auto-GPT/ and then analyze the source code and remove any unused or unnecessary code elements.', 'Provide a comprehensive documentation of any modifications made to the original code. Organize the documentation into the sections outlined in the previous response.']

[alexandruioanvoda]

This is genuine curiosity: why is this labeled as a limitation of the AI model, @richbeales ?

[Pwuts]

@alexandruioanvoda the model tries to execute a command that doesn't exist. It should know that based on the system prompt. So if it tries anyways, that's "dumb" or a "limitation of the AI model". Nothing we can really do about it.

[AayushGoyal22]

I am having the same issue please help

[estebgonza]

I had the same problem when I tried to clone a repository.
I found the solution with #530

I activated EXECUTE_LOCAL_COMMANDS=True in the .env file and now it's working.

[Pwuts]

@estebgonza if that fixes your problem it's likely a side effect, rather than an actual fix. A side effect with a relatively dangerous main effect: letting your bot execute arbitrary commands on your machine. Beware.

[Mohan33821]

it is not working to me but:
Command execute_command returned: Unknown command 'execute_command'. Please refer to the 'COMMANDS' list for available commands and only respond in
the specified JSON format.

[cutmasta-kun]

For me the same. How to add a custom Command to the COMMANDS List?

[zer0int]

@cutmasta-kun
Assuming you have edited the execute_code.py to expand it with a command, you also need to add that to app.py - that is / contains the "COMMAND" list:

from autogpt.commands.execute_code import my_function

...

    elif command_name == "my_function":
        return my_function(arguments["some_returned_info"])

[cutmasta-kun]

@zer0int
Thanks for your Feedback! Indeed, I found out, that I need to Change the exact Same files and lines you mentioned 😅
But I didn't try it yesterday 🙈
Oh! And I also added a few aliases, just to be sure 👌

[Hariel07]

I didn't understand, where what line of code do I have to put in execute_code.py and where?

[Aherontas]

@cutmasta-kun can you share with us what exactly you have done? Because what i try in my case doesn't seem to work as supposed so.

[cutmasta-kun]

@Aherontas @Hariel07 I want to try to help you:
After you cloned the repository, you have a folder in root called autogpt.
In there are the mentioned file app.py and the folder commands.
In commands are the default COMMANDS your Instance will be available to use.
You create a new file root/autogpt/commands/execute_code.py which will export the function my_function.
In root/autogpt/app.py you add the line

from autogpt.commands.execute_code import my_function

and edit the elif where the commands are listed. Just search for something in the form of:

    elif command_name == "my_function":
        return my_function(arguments["some_returned_info"])

execute_code.py and my_function are just example names.
I hope this helps you 😊👍

@zer0int I hope I explained it right ^^'

[milioe]

Maybe you should check if your .env file does not contain quotation marks on keys which eventually impede model to take actions