The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides instructions to clone the official llama.cpp repository from GitHub (ggml-org/llama.cpp) and download model weights from Hugging Face (meta-llama/Llama-3.1-8B). These are well-known and trusted sources within the AI development ecosystem.
[COMMAND_EXECUTION]: The documentation includes standard build and execution commands (make, llama-cli, llama-quantize) necessary for local model optimization. It also features Python snippets that use subprocess to check for available VRAM via nvidia-smi to guide hardware offloading settings.
[REMOTE_CODE_EXECUTION]: Instructions specify the installation of legitimate libraries like llama-cpp-python and transformers via pip, which are standard requirements for the described workflows.

gguf-quantization