Steve

CI is red due to the use of is_relative_to here https://github.com/Significant-Gravitas/Auto-GPT/blob/5802f17726e69d4a764254c4f3d6a259aa410839/autogpt/workspace.py#L38, which only exists in 3.9+ unsure how to resolve this

I tried that before, but it was rejected as someone didn't like the fact that I was updating to an older method call. Let's try again

Edit: i was wrong here, fix is to update your python

I updated my dockerfile to FROM python:3.10

Check your python interpreter and make sure it's 3.11 and not 3.8, check your dockerfile if using devcontainer python -V

Having a command which can insert into files at various byte offsets may be useful, but I'm not sure if the current LLM would use it rather than just reading...

@nalbion @pencilvania I have a PR out to remove that check, it doesn't seem to be very helpful and causes more issues than it solves.

The check is not very robust by design, appending the same content to the same file multiple times is a perfectly reasonable operation. So is writing to the same file...

> I'm not saying that `write_to_file` _should_ insert a new snippet at an offset (but I think that would be useful as another command). I'm saying that it should not...

Unable to repro this  autoGPT correctly creates a test file in this case with the right data. Are you on stable or master?