Keep comments and docs to the minimum so the code stays the single source of truth. Comments state why in three lines at most. Docs record decisions, never how the code works. AGENTS.md is the only instruction file and holds how to work, never the state of the code. Prose never points into the code. Every change fixes or deletes the prose it makes false.
安装
npx skills add skrrt-sh/skills --skill self-documentingnpx skills add skrrt-sh/skills --skill self-documenting -a claude-code -gnpx skills add skrrt-sh/skills --skill self-documenting -a codex -ggemini skills install https://github.com/skrrt-sh/skills.git --path skills/dev/self-documentinggh skill install skrrt-sh/skills self-documentinggit clone --depth 1 https://github.com/skrrt-sh/skills /tmp/skrrt-sh__skills
mkdir -p ~/.claude/skills ~/.agents/skills
cp -R /tmp/skrrt-sh__skills/skills/dev/self-documenting ~/.claude/skills/self-documenting # Claude Code
cp -R /tmp/skrrt-sh__skills/skills/dev/self-documenting ~/.agents/skills/self-documenting # Codex, Cursor, Gemini CLI, Copilot, OpenCode