Comment by dist-epoch

It is fixable.

Model requires proof that you are a legitimate developer of that piece of software.

Every Anthropic/OpenAI account will have a list of projects the model is allowed to work on for security issues.