Independent research documenting the internal prompt architecture, agent directives, and security classifiers of Claude Code - View it on GitHub