Trying to appeal to the LLM is pointless, everyone would just claim that they're working on their own code.
I've found that if you're seeing refusals and you have a more permissive model available, you can use that to find the issues and pass them back to Claude for review and implementation.
I've found that if you're seeing refusals and you have a more permissive model available, you can use that to find the issues and pass them back to Claude for review and implementation.