Anthropic appears to have accidentally revealed how one of its most important AI products works. A large internal file linked to its agentic AI tool, Claude Code, was made public. The issue happened ...