菜单
Use for investigation, review, audit, performance inspection, 調査, レビュー, 監査, or 性能確認 in doge-shell. Keeps work read-only first, avoids broad tests early, and narrows likely files before editing.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Use for doge-shell notebook, markdown rendering, output history, tm/out display, notebook_play, or terminal markdown work.
Use when working in the doge-shell repository or doge-shell リポジトリ. Routes to the right crate or narrower doge-shell skill, avoids broad reads, and chooses the smallest effective Rust validation command.
Use for doge-shell serve, static files, HTTP handlers, CORS, scanner, path validation, or web serving work.
Use for doge-shell builtin commands, proxy builtins, help, project, export, task, snippet, git builtins, or 組み込みコマンド work outside chat/MCP tools.
Use for doge-shell command palette, AI actions, diagnose, explain, suggest, argument_explainer, command_palette, ai_features, or コマンドパレット work.
Use for doge-shell Lisp, config.lisp, config loader, startup Lisp, stdlib, Lisp設定, 起動設定, include, or reload work. Narrows reads to Lisp/config startup paths.
| name | doge-shell-investigation |
| description | Use for investigation, review, audit, performance inspection, 調査, レビュー, 監査, or 性能確認 in doge-shell. Keeps work read-only first, avoids broad tests early, and narrows likely files before editing. |
rg --files or rg -n; do not edit or run broad tests first.README.md or running workspace-wide commands.