Commit

Does anyone have a good guide to running an agent using a local LLM (ollama or llama.cpp)? I can get the basic text responses running but the generating code etc just falls over and crashes