The Agent Skills Directory

[COMMAND_EXECUTION]: The skill performs extensive system profiling using platform-specific commands such as nvidia-smi, sysctl, wmic, and system_profiler to identify hardware tiers and recommend appropriate LLM models.
[COMMAND_EXECUTION]: The workflow manages the Ollama service lifecycle and environment configuration using pkill, ollama serve, and launchctl to apply performance-tuning variables like OLLAMA_FLASH_ATTENTION.
[EXTERNAL_DOWNLOADS]: The mandatory synchronization phase in SKILL.md executes git fetch and git pull to update the skill's local environment from its remote source repository before performing edits.
[DATA_EXPOSURE]: System detection scripts (detect_system.py) collect technical hardware specifications, including CPU models, VRAM capacity, and existing Ollama environment variables, to facilitate tailored optimization profiling.

ollama-optimizer