A Hugging Face contributor announced the adapter on r/LocalLLaMA and described it as part of the group’s work on adapters trained for specific agent harnesses. The post includes commands to download the model and adapter, serve them with llama.cpp, and install the Pi integration.