feat: add bubseek langchain

[PsiACE]

We integrate with the LangChain ecosystem by treating Runnable as the single execution contract. In LangChain, a Runnable is a composable unit of work that supports synchronous and asynchronous invocation, batching, streaming, transformation, and composition, with config carrying metadata and tags for tracing. This makes it the right abstraction for plugging Bub into LangChain, LangGraph, DeepAgents, and other LangChain-based runtimes through one stable interface (https://reference.langchain.com/python/langchain-core/runnables/base/Runnable))

Our approach is to adapt Bub’s run_model hook to accept any LangChain Runnable (or a factory that returns one), pass execution context through Runnable metadata, and bridge Bub-native tools into LangChain-compatible tools so they can be discovered and called by agents. On top of that, we expose Bub’s skills, tools, and tape-aware observability in a way that feels native to LangChain, while keeping Bub as the runtime and orchestration layer.