菜单
Comprehensive codebase research that scales effort to task complexity, synthesises findings, and writes a timestamped document to docs/research/. Use when user says /research-codebase, asks to research or document a codebase area, or wants a deep-dive into how something works.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
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。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/damian87x/oh-my-copilot --skill research-codebase命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Split an approved plan into parallel tmux panes in the current window so the user can watch agents work. Prefer this visual flow by default; use `omp team` only when the user explicitly wants background execution or runtime messaging/status APIs.
Gate-driven retry loop — re-run an agent until a shell gate command passes (e.g. PR checks green, tests passing). Use with /loop when a machine-checked condition decides done-ness, not the model's say-so.
Use when the user invokes /goal, wants a durable objective for the current Copilot session, asks to resume an unfinished objective, or needs turn-based continuation with completion and blocker decisions.
基于 SOC 职业分类
正在显示 SKILL.md
| name | research-codebase |
| description | Comprehensive codebase research that scales effort to task complexity, synthesises findings, and writes a timestamped document to docs/research/. Use when user says /research-codebase, asks to research or document a codebase area, or wants a deep-dive into how something works. |
| argument-hint | <area or question to research> |
Read-only codebase research. Documents what EXISTS — no recommendations or critiques unless explicitly asked.
Respond: "I'm ready to research the codebase. What area or question should I investigate?" — then wait.
If the user references specific files, read them directly in the main context before anything else.
Detect from package.json, config files, directory structure, and project conventions. This determines which reference files are relevant — do NOT load unrelated framework guidance.
| Tier | When | Agent strategy |
|---|---|---|
| Small | Single file/component, narrow question | Read and search directly (glob/grep/read). No delegation. |
| Medium | Cross-file, single area (e.g. "how does auth work") | Locate with glob/grep, then read and analyse the hits directly. |
| Large | Cross-cutting, multi-area (e.g. "map the entire API layer") | Sweep area-by-area (locate → analyse → find patterns), or delegate areas to parallel omp team workers (see reference/agent-prompts.md). |
reference/agent-prompts.md for omp team worker prompt templatesreference/template.md for output format and frontmatterreference/permalink.md if GitHub permalinks are applicabledate, user, commit, branch, repodocs/research/YYYY-MM-DD-description.mdShow concise summary to user with key file references. Ask if they have follow-up questions.
If the user has follow-ups, read reference/follow-up.md for the append protocol.
This skill can drive multiple tool calls or long-running output. Use omp cost [--today] [--session <id>] for local hook-ledger estimates only; it is not provider billing. Keep injected summaries concise and prefer bounded output when rerunning noisy commands.