SafetyHugging Face Published the OpenAI Agent's Commands. Anthropic's Models Wouldn't Read ThemJul 30, 6:16 AM