purevocals-uvr-automator

This skill is coherent with its stated purpose, but it will perform large runtime operations on your machine: it will create/activate a virtual environment, install PyTorch (GBs), install audio-separator and other Python packages via pip (using a mirror), download pretrained model files and ffmpeg binaries, and run subprocesses (including nvidia-smi) to detect GPU. It does not request credentials, but it will write files under the skill directory (models, venv, logs) and any input/output paths you give it. Before installing or invoking: 1) Ensure you trust the author/sources; 2) Expect substantial network and disk usage and long installs; 3) If you want to limit risk, run the script manually in an isolated environment or container (so installs are contained) or inspect/modify VENV_DIR to use a per-skill venv instead of a shared one; 4) Avoid pointing the skill at folders containing sensitive files; 5) If you need stricter control, run the code locally yourself and review the model download URLs and pip install commands. Overall the behavior is expected for a local model-based audio tool, but treat the automatic installs and shared venv as non-trivial operational changes and proceed accordingly.