Anthropic accidentally released parts of its Claude Code AI agent, sparking industry-wide debate on AI security, transparency, and the risks of operational errors in fast-growing tech ecosystems.