They have an interesting regex for detecting negative sentiment in users prompt which is then logged (explicit content): https://github.com/chatgptprojects/claude-code/blob/642c7f94... I guess these words are to be avoided...
No discussion yet. Be the first to share your thoughts!