field-books-survey

Field Survey — One-Click Comprehensive Discipline Research

Enter a discipline name to automatically research the discipline classification system, collect authoritative book lists for each direction in parallel, and output a structured classification document and subdirectory skeleton.

When to Use

Triggered when the user requests the following operations:

"调研心理学" / "调研经济学"
"心理学学科调研" / "法学学科调研"
"survey psychology" / "survey computer science"
"研究生物学有哪些方向" / "计算机科学有哪些分支"
"调研心理学-临床方向" (complete classification + only in-depth clinical direction)

Phase 0: Parse Input

Based on the user input, identify the following information:

Discipline Name(required)
- Extract the discipline name from the user message
- "调研心理学" → Psychology
- "survey computer science" → Computer Science
Focus Direction(optional)
- "调研心理学-临床方向" → Focus Direction = Clinical Psychology
- If unspecified, it is a comprehensive survey, and all directions are collected in depth
Whether to Skip In-Depth Book Analysis
- Default:Automatically execute Phase 4 (call analyze-book for in-depth analysis of each book)
- Skip Phase 4 only when the user explicitly says "仅分类"/"不分析"/"跳过分析"/"只要书单"
- {skip_analysis}: true/false
Determine Key Variables(needed in all subsequent phases)
- {discipline}: Chinese name of the discipline (e.g., "心理学")
- {discipline_en}: English name of the discipline, lowercase hyphen-separated (e.g., "psychology")
- {focus}: Focus direction name (optional, e.g., "临床心理学"; empty if unspecified)
- {date}: Current date, in YYYY-MM-DD format
- {skip_analysis}: Whether to skip Phase 4 (default false)
Output Directory: Current working directory

If the discipline name or its corresponding English name cannot be determined, first use one WebSearch for a quick lookup,do not skip this step。

Phase 1: Sort Out the Discipline Classification System (main thread)

Comprehensively sort out the discipline classification system through the four-dimensional information source framework.

Four-Dimensional Information Source Framework

Channel	Information Source Type	Search Strategy	Priority
Education System	Ministry of Education discipline catalog, university department settings, course syllabi	`"{discipline}" 学科目录教育部` / `"{discipline}" 院系设置课程`	Highest
Academic Organizations	Society/association branch settings, research direction divisions	`"{discipline}" 学会分会研究方向` / `"{discipline}" association divisions`	High
Practice Institutions	Hospital departments, research institutes, law firm departments, court divisions	`"{discipline}" 科室设置` / `"{discipline}" 研究所部门` / `"{discipline}" 业务方向`	High
Industry/ Sector	Professional qualification certification, industry associations, career paths	`"{discipline}" 职业资格认证` / `"{discipline}" 行业分类`	Medium

Discipline-Institution Mapping Examples

The following are the four-dimensional information source correspondences for common disciplines:

Discipline	Education System	Academic organizations	Practitioner institutions	Industry/enterprise
Medicine	Medical schools	Chinese Medical Association	Hospital departments	Health Commission
Law	Law schools	China Law Society	Courts/Law firms	Judicial examination
Psychology	Departments of Psychology	CPA/APA	Hospital psychology departments	Counselor certification
Biology	Colleges of Life Sciences	Biological societies	CAS/NIH	Biotechnology
Economics	Schools of Economics	Economics societies	Central Bank/NDRC	Financial industry
Computer Science	CS schools	CCF/ACM/IEEE	Technology research institutes	Internet industry

Execution requirements

Execute a total of 5-8 WebSearch + 2-3 WebFetch, covering four channels
Search order: first the education system (the most authoritative classification standard), then academic organizations and practitioner institutions (supplementation and verification), and finally industry/enterprise (supplementary application perspective)
WebFetch priority access: the official page of the Ministry of Education discipline catalog, the branch list on the official website of authoritative societies
Cross-validation rules：
- Mentioned by ≥2 channels = mainstream direction, marked as "mainstream"
- Mentioned by only 1 channel = emerging/niche direction, marked as "emerging"
- Education + academic organizations = academic core direction; practice + industry but no education = application-oriented direction
Organize each direction as: number, direction name, one-sentence description, research object/practice field
Expected total number of directions is 8-15; if more than 15, merge highly related subdirections

Output format

Organize the sorting results into a structured direction list:

方向 N: {方向名称}（{英文名}）
- 核心研究对象：{描述}
- 实践领域：{描述}
- 信息源：{哪些渠道提及}
- 分类：主流方向 / 新兴方向

Focused direction handling: if the user specified {focus}, still output the complete classification system, but only mark {focus} the corresponding directions to enter Phase 2 in-depth collection. If not specified, all directions enter Phase 2.

Phase 2: Parallel booklist collection (N Subagents)

Key requirement: Agents for all directions must be placed inthe same messageto ensure parallel execution.

Each Agent uses subagent_type: "general-purpose"。

Dispatch 1 subagent per direction to collect booklists according to the following 4 channels (in decreasing priority):

Priority	Channel	Collection target	Search strategy
1 (core)	Academic channel	Textbooks, course booklists, foundational works	`"{direction_name}" 教材推荐` / `"{direction_name}" textbook recommended syllabus`
2 (important)	Practitioner institutions	Clinical guidelines, operation manuals, institutional training textbooks	`"{direction_name}" 临床指南手册` / `"{direction_name}" 实务操作培训教材` / `"{direction_name}" 实践 handbook`
3 (supplementary)	Industry channel	Industry association recommendations, practitioner exam reference books	`"{direction_name}" 从业者推荐书` / `"{direction_name}" 必读书目入门`
4 (verification)	General/popular	Douban/Goodreads ratings, booklist recommendations	`"{direction_name}" 豆瓣书单` / `"{direction_name}" 好书推荐评分`

Language strategy: For disciplines with a high degree of internationalization (computer science, linguistics, economics, etc.), use both Chinese and English search keywords. For discipline directions with strong Chinese characteristics (traditional Chinese medicine, Chinese law, etc.), prioritize searching in Chinese.

Sub-agent Prompt Template

Each direction's sub-agent uses the following complete prompt:

description: "Collect a book list for the {discipline} · {direction_name} direction"

prompt template:

你是学术书籍调研专家。请为「{discipline}」学科的「{direction_name}」方向采集权威书单。

方向编号：{direction_number}
方向描述：{direction_description}

任务：按以下 4 个渠道依次采集，至少执行 3 次 WebSearch + 1 次 WebFetch：

**渠道 1 — 学术渠道（核心）**
搜索关键词：
- "{direction_name}" 教材 推荐
- "{direction_name}" textbook recommended syllabus
- "{direction_name}" 经典教材 课程书单
目标：找到该方向的标准教材、课程指定用书、奠基性著作。

**渠道 2 — 实践机构（重要）**
搜索关键词：
- "{direction_name}" 临床指南 手册
- "{direction_name}" 实务操作 培训教材
- "{direction_name}" 实践 handbook
目标：找到实践指南、操作手册、机构培训用书。

**渠道 3 — 行业渠道（补充）**
搜索关键词：
- "{direction_name}" 从业者 推荐书
- "{direction_name}" 必读书目 入门
目标：找到行业协会推荐书、职业考试参考书、从业者必读。

**渠道 4 — 大众/通俗（验证）**
搜索关键词：
- "{direction_name}" 豆瓣 书单
- "{direction_name}" 好书推荐 评分
目标：通过豆瓣/Goodreads评分验证书籍质量，补充高评分通俗读物。

使用 WebFetch 访问至少 1 个搜索结果页面，验证书籍信息的准确性。

输出格式（严格遵守）：

**{direction_number}. {direction_name} 书单**

| 书名（中文） | 英文原名 | 作者 | 出版社 | 版次/年份 | 简介 | 推荐理由 |
|-------------|---------|------|--------|----------|------|---------|
| {中文书名} | {英文名或N/A} | {作者} | {出版社} | {版次/年份} | {一句话简介} | {为何推荐+来源渠道} |

**来源链接：**
- [{来源标题}]({url})
- ...

输出要求：
- 5-10 本书，按优先级排序：教材/奠基之作 > 经典专著 > 通俗高评分
- 总输出控制在 1200 字符以内
- 全部使用中文输出
- 来源链接必须为实际访问过的 URL，不得使用占位符
- 如某渠道无结果，注明"该渠道未找到相关书籍"并跳过

Sub-agent Constraints

Each sub-agent outputs 5-10 books
Output character limit 1200 characters
All output in Chinese
Must cite source URLs
Order: textbooks/foundational works > classic monographs > popular high-rated

Phase 3: Integration and Output (Main Thread)

After waiting for all sub-agents to return, complete the following steps in the main thread.

Step 1: Data Aggregation

Consistency check: the same book appears in multiple directions → mark as "cross-domain classic" and mention it specifically in the core conclusions
Cross-direction deduplication: keep the same book in the most relevant direction (judged by the recommendation reason), and reference it in other directions as "see XX direction"
Source completeness: check whether the source URLs for all directions are complete; mark missing ones as "source to be supplemented"
Sub-agent failure handling: mark directions that did not return results as "insufficient data, manual supplementation recommended"
Statistical summary: calculate the total number of books and the number of cross-domain classics, for use in the core conclusions and terminal summary

Step 2: Generate the main classification document

File name:{discipline_en}-classifications-{date}.md, saved in the current working directory.

Use the Write tool to write the following Markdown content:

# {discipline}分类全景

> 调研日期：{date}
> 文档目标：多角度全面梳理{discipline}的学科体系、实践机构设置、权威书籍分类

## 核心结论

{基于四维调研得出的 3-5 条核心发现，包括：主流/新兴方向数量、最具交叉性的方向、跨领域经典书籍数量、教育与行业分类差异、推荐入门路径}

## 学科地图总览

| 层级/大类 | 典型分支 | 关注重点 |
|-----------|---------|---------|
| {大类1} | {分支1}, {分支2}, ... | {关注重点描述} |
| {大类2} | {分支3}, {分支4}, ... | {关注重点描述} |
| ... | ... | ... |

## {编号}. {方向名称}（{英文名}）

**核心研究对象**：{该方向的研究对象描述}

**实践领域**：{该方向的实践应用领域}

**研究方法**：{该方向常用的研究方法}

**代表性书籍：**

| 书名（中文） | 英文原名 | 作者 | 出版社 | 版次/年份 | 简介 | 推荐理由 |
|-------------|---------|------|--------|----------|------|---------|
| {书名} | {英文名} | {作者} | {出版社} | {版次/年份} | {简介} | {推荐理由} |
| ... | ... | ... | ... | ... | ... | ... |

{对每个方向重复上述区块}

## 参考来源

**教育体系：**
- [{来源标题}]({url})

**学术组织：**
- [{来源标题}]({url})

**实践机构：**
- [{来源标题}]({url})

**行业/产业：**
- [{来源标题}]({url})

**各方向书单：**
- {方向1}: [{来源标题}]({url}), ...
- {方向2}: [{来源标题}]({url}), ...
- ...

## 下一步

本文档为学科全景调研结果，每个方向的书籍已整理在对应子目录中。如需对某本书进行深度学习，可使用 `/analyze-book` 技能：

1. 进入感兴趣的方向子目录
2. 查看该方向的 README.md 了解书单
3. 对目标书籍执行 `/analyze-book {书名}` 生成深度学习笔记

建议优先从"核心结论"中推荐的入门路径开始，逐步扩展到其他方向。

Markdown filling rules:

Replace the placeholders in the template with actual data
The core conclusions are written based on the results of the four-dimensional cross-validation and must be specific and actionable
Cross-domain classic books are listed in detail in the direction where they first appear, and other directions note "see XX direction"
The "References" section is required: fill in all source links from Phase 1 and Phase 2, grouped by category; do not omit them and do not use placeholder URLs

Step 3: Create the subdirectory skeleton

Use the Bash tool to create the directory structure.

Naming rules：

By default, use two-digit sequential numbering:01-{direction_en}, 02-{direction_en}, ...
If the discipline has a clear higher-level classification hierarchy (such as internal medicine/surgery in medicine), use hierarchical numbering:1-{category}/1.1-{direction_en}, ...
{direction_en} Use lowercase hyphen-separated English names
Example (psychology, default numbering):01-clinical-psychology/, 02-cognitive-psychology/, ...
Example (medicine, hierarchical numbering):1-internal-medicine/1.1-cardiology/, 2-surgery/2.1-general-surgery/, ...

Create in each subdirectory README.md, and use the Write tool to write the following content:

# {方向名称}（{英文名}）

> 所属学科：{discipline}
> 调研日期：{date}

## 方向简介

{方向描述，包含核心研究对象和实践领域}

## 书单

| 书名（中文） | 英文原名 | 作者 | 出版社 | 版次/年份 | 简介 | 推荐理由 | 状态 |
|-------------|---------|------|--------|----------|------|---------|------|
| {书名} | {英文名} | {作者} | {出版社} | {版次/年份} | {简介} | {推荐理由} | 待分析 |
| ... | ... | ... | ... | ... | ... | ... | 待分析 |

## 深度学习

如需对某本书进行深度学习，可使用：
- `/analyze-book {书名}` — 生成深度学习笔记

Step 4: Output progress information in the terminal, then immediately continue

Print the following progress information in the terminal:

📋 Phase 1-3 完成：{discipline} 分类与书单已生成

📄 主文档：{discipline_en}-classifications-{date}.md
📁 子目录：{N} 个方向目录已创建
📚 书籍总计：{总书籍数} 本（跨领域经典 {N} 本）

目录结构：
  {列出所有子目录名称}

⚠️ Note: after printing the above progress information, the task is not yet complete.

If {skip_analysis} is true: print "ℹ️ Skipped in-depth book analysis (user specified classification only)", then end.
Otherwise:Immediately begin executing Phase 4, and do not stop.

Phase 4: Automatic In-Depth Book Analysis (Executed by Default)

Skip this phase only when the user explicitly says "classify only"/"no analysis"/"skip analysis"/"book list only". By default, after Phase 3 is completed, Phase 4 is entered automatically.

Execution strategy

By directionone by oneprocessing (to avoid launching too many subagents at the same time), and within each direction, books arein parallelanalyzed:

方向1: 书A ──┐
       书B ──┼── 并行 → 全部完成 → 方向2: 书C ──┐
       书C ──┘                            书D ──┼── 并行 → ...
                                          书E ──┘

If the user specifies a focus direction {focus}, then execute Phase 4 only for that direction.

Processing for each direction

For all books in direction M (assuming there are K books), inthe same message, issue K Agent calls: