This is a comprehensive workflow for designing therapeutic proteins from scratch using AI tools like RFdiffusion for backbone generation and ProteinMPNN for sequence optimization. You'll go through target characterization, generate multiple backbone candidates, design sequences, validate structures with ESMFold or AlphaFold2, and assess developability metrics like aggregation and expression potential. The skill enforces a report-first approach and includes detailed checklists to keep you from skipping validation steps. It's built around NVIDIA NIM tools, so you'll need an API key from build.nvidia.com. The documentation is thorough about parameter naming quirks and common mistakes, which matters when you're running expensive design cycles. Best for computational biologists who need structured guidance through the entire de novo protein design pipeline.
npx skills add https://github.com/mims-harvard/tooluniverse --skill tooluniverse-protein-therapeutic-design