Description
Core abstractions for LLM applications can define messages, runnables, callbacks, prompts, and model interfaces. Developers use this package as the shared foundation for LangChain-based systems. Even core layers can carry prompt text, metadata, and callback traces that need privacy review.