This is the conservative environment setup stage for reproducing deep learning repos from their READMEs. It translates setup instructions into conda environments, figures out where checkpoints and datasets should live, and surfaces any missing dependencies before you try to run anything. Use it after you've picked a reproduction target but before you start executing training or inference commands. It won't select targets for you, won't interpret papers, and won't report final results. The scope is narrow by design: prepare the environment and asset paths, flag what's missing, then hand off. If your repo already has a working environment or you're just doing generic conda troubleshooting, skip this.
npx -y skills add lllllllama/rigorpilot-skills --skill env-and-assets-bootstrap --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
juliusbrussee/caveman
mattpocock/skills
obra/superpowers
forrestchang/andrej-karpathy-skills
vercel-labs/skills