Deep learning projects often rely on specific environments and files. Before running any code, you need to set up a conda-first environment and find the right checkpoints and datasets.
The env-and-assets-bootstrap skill provides a careful and conservative approach. It creates setup notes, candidate conda commands, and an asset path plan. It also lists any missing dependencies or risks.
This skill helps you avoid common setup mistakes. It works best after you have identified a repository to reproduce.
Global
mkdir -p ~/.claude/skills/env-and-assets-bootstrapProject
mkdir -p .claude/skills/env-and-assets-bootstrapSource Repository