"Full Installer" with pre-packaged llama.cpp

[dan-menlo]

Goal

• Cortex.cpp's current installer is not suited for air-gapped use cases, as it downloads llama.cpp at install time
• We should offer a "Full Installation" option, which has llama.cpp pre-packaged

Out-of-scope

• Cortex.cpp allows users to create a "custom installer" that bundles engines and models