Example run: Why is it rewriting history?

[canadaduane]

Ok, so I have cloi working! I tried /history and ran a previous command that I'd run in my zsh shell yesterday:

lsof -n -P -i | grep 3000

The result was a bit confusing (see screenshot below).

1. When I run this command in my shell, it seems to run just fine: no processes are running on port 3000, so it exits (with code 1)
2. When cloi runs it, I guess it detects that it's a command with a failing exit code? I don't need it to do anything, I was just curious what would happen.
3. Then it tries to fix something...
4. And it suggests rewriting history.

I don't understand what "rewriting history" means here, or why it is desirable.

Thanks!

[Kreijstal]

im not him, but maybe history contains commands that are successful, which is useful to execute them again.. rewriting history in the sense that history now contains valid useful commands rather than commands that didn´t work... just a hunch

[emirbartu]

im not him, but maybe history contains commands that are successful, which is useful to execute them again.. rewriting history in the sense that history now contains valid useful commands rather than commands that didn´t work... just a hunch

Seems like it is just considering most of things as "threat" and tries to fix it.

I just had a cursory look at the project. But there are two solutions that come directly to my mind:

1. Prompt engineering to make a more conscious model.
2. Adding logic for some basic commands or errors to not act as a "threat." For instance, if the command involved grep and the final exit code was 1, the program could recognize this as "no matches found" and simply report that to the user.