The Agent Skills Directory

[SAFE]: The skill is a collection of documentation files providing guidance on running LLM inference. No suspicious logic, obfuscation, or data exfiltration patterns were detected in the instructions.
[EXTERNAL_DOWNLOADS]: The skill instructs the user or agent to clone the llama.cpp source code from its official GitHub repository (ggerganov/llama.cpp) and download quantized models from Hugging Face. These are established and well-known services within the AI ecosystem.
[COMMAND_EXECUTION]: The documentation provides standard CLI instructions for building the C++ project and running the resulting binaries. These commands are consistent with the intended purpose of setting up a local inference environment.

llama-cpp