Teaches Claude the constraints that make CLIs feel right to users: errors go to stderr so piping works, CLI args beat env vars beat config files, non-zero exits for automation, and progress bars for anything slow. Shows the clap derive patterns for argument parsing, how to layer configuration sources properly, and which crates handle what (indicatif for progress, dialoguer for prompts, ratatui for full TUIs). The trace down sections connect domain rules to actual Rust implementations, which is helpful when you're trying to figure out why your CLI feels janky. Honestly just a solid reference for the stuff you always forget between CLI projects.
npx -y skills add zhanghandong/rust-skills --skill domain-cli --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
wshobson/agents