system-study

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

System Study — 系统化学习材料生成

System Study — Systematic Learning Material Generation

你要学一个新领域。AI 自主调研 + 搭体系 + 产出 HTML 学习材料。

和已有 skill 的本质区别：

vs long-research：那个是"有困惑 → 求结论"；这个是"想学 → 求体系"
vs case-radar：那个给散点案例集；这个给系统化全貌
vs readable-output：那个处理已有素材；这个自己挖素材 + 自己建体系（最后可调用它做出稿，但不是核心）

这个 skill 的差异化命门：除了"骨架 + 案例 + 工程化"这些教科书必讲的部分，强制产出"争议焦点 + 盲区清单"两块——这是普通学习材料没有的，是用户研究 → 内容/产品的金矿。

You want to learn a new field. AI conducts independent research + builds a knowledge system + generates HTML learning materials.

Essential Differences from Existing Skills:

vs long-research: That one is "have confusion → seek conclusions"; this one is "want to learn → seek a systematic framework"
vs case-radar: That one provides scattered case collections; this one delivers a systematic full picture
vs readable-output: That one processes existing materials; this one independently gathers materials + builds a system (it can call readable-output for final drafting, but that's not the core)

Differentiating Core of This Skill: In addition to textbook-standard sections like "framework + cases + engineering", it is mandatory to produce two sections: "Controversy Focuses + Blind Spot List" — these are not found in ordinary learning materials, and are goldmines for user research → content/product development.

5 阶段流程总览

Overview of 5-Stage Process

阶段	名称	介入次数	产物	详细操作
1	题目收敛	1 次（必）	`01-题目确认.md`	reference/stage-1-topic-narrow.md
2	调研计划	1 次（默认）	`02-调研计划.md`	reference/stage-2-research-plan.md
3	Sub-agent 并行调研	0 次	`03-调研笔记/A-E.md`	reference/stage-3-subagent-templates.md ⭐
4	体系搭建+大纲确认	1 次（默认）	`04-体系大纲.md`	reference/stage-4-synthesis.md
5	HTML 学习材料	0 次	`05-学习材料.html`	reference/stage-5-html-spec.md

默认严格逻辑：阶段 1/2/4 都问用户。但用户可一次性说"全程不打扰、直接出 HTML"，跳过 2、4。阶段 1 永远不能省——题目错了后面全白干。

Stage	Name	Number of Interventions	Deliverables	Detailed Operations
1	Topic Narrowing	1 time (mandatory)	`01-题目确认.md`	reference/stage-1-topic-narrow.md
2	Research Plan	1 time (default)	`02-调研计划.md`	reference/stage-2-research-plan.md
3	Sub-agent Parallel Research	0 times	`03-调研笔记/A-E.md`	reference/stage-3-subagent-templates.md ⭐
4	System Building + Outline Confirmation	1 time (default)	`04-体系大纲.md`	reference/stage-4-synthesis.md
5	HTML Learning Materials	0 times	`05-学习材料.html`	reference/stage-5-html-spec.md

Default Strict Logic: The system will consult the user in Stages 1/2/4. However, the user can say "Don't disturb me throughout, directly output HTML" to skip Stages 2 and 4. Stage 1 can never be skipped — if the topic is wrong, all subsequent work will be in vain.

用户授权"全程不打扰"快速通道

Fast Track with User Authorization of "No Disturbance Throughout"

如果用户在阶段 1 / 2 / 任何阶段说出"你直接来吧 / 直接给我结果就行 / 你自己定 / 你判断 / 全程不打扰"——主线程立即切换到快速通道：

阶段 1 依然要问（用 AskUserQuestion 弹切片）——题目永远不能 AI 替用户拍
阶段 2 仍然写
02-调研计划.md
存档，但不再问用户，直接进阶段 3
阶段 4 仍然写
04-体系大纲.md
存档，但不再问用户，直接进阶段 5
在 PLAN.md 标注"用户已授权全程不打扰"——这是审计线索

切换条件：用户的话里明确包含"直接 / 不打扰 / 自己定 / 自己判断 / 你来"等授权词。模糊表达（如"看着办"）不算——继续按默认严格流程问。

If the user says phrases like "Just do it / Just give me the result / You decide / You judge / No disturbance throughout" in Stage 1 / 2 / any stage — the main thread immediately switches to the fast track:

Stage 1 still requires consultation (use AskUserQuestion to send options) — the topic can never be decided by AI on behalf of the user
Stage 2 still writes
02-调研计划.md
for archiving, but no longer consults the user, directly proceeds to Stage 3
Stage 4 still writes
04-体系大纲.md
for archiving, but no longer consults the user, directly proceeds to Stage 5
Mark "User authorized no disturbance throughout" in PLAN.md — this is an audit trail

Switching Condition: The user's words explicitly contain authorization terms like "directly / no disturbance / you decide / you judge / you do it". Vague expressions (e.g., "do as you see fit") do not count — continue with the default strict process.

核心设计原则（5 条）

Core Design Principles (5 Rules)

这 5 条是从首发实战提炼的、贯穿整个 skill 的设计哲学。所有 stage 都遵循：

必经卡点不省——阶段 1 题目收敛永远不能省。其他卡点用户可授权跳过，但题目错了全白干。
强制要异见，禁止和稀泥——Agent E 不接受"两边都对"。每个议题必须给 X/10 倾向分。
盲区清单独立成块——还没共识的、还在打架的、AI 互相复读的，独立段落，不混在其他内容里。
用户研究视角主导——不走"基础→进阶→高阶"教材路线，按"我能拿来做什么"组织。允许带判断、带视角，不需要装客观。
HTML 独立写，不调 readable-output——学习材料需要侧栏导航 + 章节折叠 + 争议高亮 + 盲区独立区，长文阅读结构不够用。

These 5 rules are refined from initial practical applications and are the design philosophy that runs through the entire skill. All stages must follow them:

Mandatory checkpoints cannot be skipped — Stage 1 topic narrowing can never be skipped. Other checkpoints can be skipped with user authorization, but a wrong topic makes all work useless.
Force dissenting opinions, avoid vague compromises — Agent E does not accept "both sides are right". Each issue must be given an X/10 inclination score.
Blind Spot List is an independent section — Content without consensus, still controversial, or repeatedly paraphrased by AI must be in an independent paragraph, not mixed with other content.
User research perspective-led — Do not follow the "basic → intermediate → advanced" textbook route; organize content around "what I can do with this". Judgments and perspectives are allowed, no need to pretend to be objective.
HTML is written independently, do not call readable-output — Learning materials require sidebar navigation + chapter folding + controversy highlighting + independent blind spot section, which cannot be satisfied by long-text reading structures.

5 个 Sub-agent 阵容（A-D 灵活 + E 必选）

5 Sub-agent Lineup (A-D Flexible + E Mandatory)

详细 prompt 模板见 reference/stage-3-subagent-templates.md。

Agent	固定职责	主题适配方式
A 骨架	官方文档 + 学派 / 主流方法	视主题决定具体源（提示词→Anthropic/OpenAI 文档；记忆系统→各家 memory 文档；MCP→协议文档+实现）
B 前沿	新形态、新趋势、最新讨论	视主题灵活——一句话定义"这领域最新的是什么"
C 真物	顶级案例 / 真实代码 / 原文拆解	必须看原文；案例数量按可信源能找到几个定
D 生态	工具景观 / 工程化 / 应用场景	直接对接用户的产品决策
E 争议+盲区	必选，强制双轨陈述 + 给倾向分 + 盲区清单	差异化命门——和市面学习材料的核心区别

主线程拿到题目后，根据主题情况微调每个 sub-agent 的具体调研范围（不是改职责类型，是改"查哪些源"），再批量派出。

Detailed prompt templates can be found in reference/stage-3-subagent-templates.md.

Agent	Fixed Responsibilities	Topic Adaptation Method
A Framework	Official documents + schools of thought / mainstream methods	Determine specific sources based on the topic (prompt engineering → Anthropic/OpenAI documents; memory systems → various memory documents; MCP → protocol documents + implementations)
B Cutting-edge	New forms, new trends, latest discussions	Flexible based on the topic — define "what's the latest in this field" in one sentence
C Authentic Cases	Top-tier cases / real code / original text breakdown	Must refer to original texts; number of cases depends on how many can be found from credible sources
D Ecosystem	Tool landscape / engineering / application scenarios	Directly align with the user's product decisions
E Controversies + Blind Spots	Mandatory, forced dual-track presentation + inclination score + blind spot list	Differentiating core — the key difference from mainstream learning materials

After the main thread receives the topic, fine-tune the specific research scope of each sub-agent based on the topic (not changing the responsibility type, but changing "which sources to check"), then dispatch all in batches.

工作流（主线程要做什么）

Workflow (What the Main Thread Needs to Do)

Step 0：建工作目录

Step 0: Create Working Directory

<主题>学习/                  # 默认目录名，可改
├── PLAN.md                  # 总控
├── 01-题目确认.md
├── 02-调研计划.md
├── 03-调研笔记/
│   ├── A-{骨架切片名}.md
│   ├── B-{前沿切片名}.md
│   ├── C-{真物切片名}.md
│   ├── D-{生态切片名}.md
│   └── E-争议与盲区.md
├── 04-体系大纲.md
└── 05-学习材料.html

<Topic>Learning/                  # Default directory name, can be modified
├── PLAN.md                  # Master control
├── 01-题目确认.md
├── 02-调研计划.md
├── 03-调研笔记/
│   ├── A-{Framework Slice Name}.md
│   ├── B-{Cutting-edge Slice Name}.md
│   ├── C-{Authentic Case Slice Name}.md
│   ├── D-{Ecosystem Slice Name}.md
│   └── E-争议与盲区.md
├── 04-体系大纲.md
└── 05-学习材料.html

Step 1：题目收敛

Step 1: Topic Narrowing

用
```
AskUserQuestion
```
（multiSelect: true）给用户 4 个切片选项让她勾
切片描述要具体到"覆盖什么"，不要泛泛
用户全选也行（这次提示词学习就是 4 个全选）
写
```
01-题目确认.md
```
记录主题 + 学习视角

Use
```
AskUserQuestion
```
(multiSelect: true) to provide the user with 4 slice options to check
Slice descriptions should be specific to "what is covered", not vague
The user can select all 4 (as in this prompt learning case)
Write
```
01-题目确认.md
```
to record the topic + learning perspective

Step 2：调研计划

Step 2: Research Plan

列出每个 sub-agent 要查的源（官方文档、社区、案例库、争议议题）
给用户审一遍，可砍可加
用户说"开跑"就进入 Step 3
写
```
02-调研计划.md
```

List the sources each sub-agent needs to check (official documents, communities, case libraries, controversial topics)
Let the user review, they can remove or add items
Proceed to Step 3 when the user says "Start"
Write
```
02-调研计划.md
```

Step 3：派 5 个 Sub-agent 并行

Step 3: Dispatch 5 Sub-agents in Parallel

每个 sub-agent 用一个独立的 Agent 调用（subagent_type: general-purpose）
一次性在同一个 message 里发出全部 5 个（并行）
prompt 模板见 reference/stage-3-subagent-templates.md
必须包含的硬约束（见 anti-patterns.md）：
- 分段写入文件（每完成 1-2 节就 Write 一次）
- 反 AI 文废话清单
- 来源可追溯
- 争议必须双轨 + 给倾向

Each sub-agent is called via an independent Agent (subagent_type: general-purpose)
Dispatch all 5 at once in the same message (parallel execution)
Prompt templates can be found in reference/stage-3-subagent-templates.md
Mandatory hard constraints (see anti-patterns.md):
- Write to files in segments (Write once after completing 1-2 sections)
- Anti-AI fluff list
- Traceable sources
- Controversies must have dual-track presentation + inclination score

Step 4：合成 + 大纲确认

Step 4: Synthesis + Outline Confirmation

读全部 5 份调研笔记
设计 8 篇章节结构（导读 / 骨架 / 实战 / 进阶 / 案例 / 工程化 / 争议 / 盲区 / 附录）
写
```
04-体系大纲.md
```
给用户过一遍坐标系

Read all 5 research notes
Design an 8-chapter structure (Introduction / Framework / Practical Application / Advanced / Cases / Engineering / Controversies / Blind Spots / Appendix)
Write
```
04-体系大纲.md
```
Let the user review the structure

Step 5：HTML 产出

Step 5: HTML Output

按 stage-5-html-spec.md 的设计 token 和组件规范
单文件 HTML，所有 CSS 内嵌
必须包含：左侧固定导航 / 4 种 callout / 案例 details 折叠 / 争议带倾向分 / 盲区清单
一次写完，不要分多次 Write（防止截断时被中间状态打断）

Follow the design tokens and component specifications in stage-5-html-spec.md
Single-file HTML with all CSS embedded
Must include: fixed left sidebar navigation / 4 types of callouts / collapsible case details / controversies with inclination scores / blind spot list
Write in one go, do not split into multiple Write operations (to avoid interruption by intermediate states during truncation)

关键硬约束（一定要看 anti-patterns.md）

Key Hard Constraints (Be Sure to Read anti-patterns.md)

Sub-agent 必须分段写入文件——不能等最后一次性写。socket 崩溃会丢全部内容。
Agent E 强制双轨 + 给倾向分——"两边都对"不接受。
案例类 Agent 强制 3+ 个原文片段 + 可信度评级——AI 总结的二手案例不接受。
AI 文废话黑名单——明显套话密度高的中文博客跳过。
主线程做合成，不下放给 sub-agent——sub-agent 只产物料。
PLAN.md 主控——每次状态变化更新进度表。

Sub-agents must write to files in segments — Do not wait until the end to write all at once. Socket crashes will cause loss of all content.
Agent E must have dual-track presentation + inclination score — "Both sides are right" is not acceptable.
Case-focused agents must include 3+ original text fragments + credibility ratings — Second-hand cases summarized by AI are not acceptable.
AI fluff blacklist — Skip Chinese blogs with high density of obvious clichés.
The main thread handles synthesis, do not delegate to sub-agents — Sub-agents only produce materials.
PLAN.md as master control — Update the progress chart every time the status changes.

灵活参数

Flexible Parameters

参数	默认	可调
切片数量	4（A-D）+ 1（E）= 5	3-6（E 必含）
介入级别	阶段 1/2/4 各一次	用户可授权全跳过 2 和 4
产物目录名	`<主题>学习/`	用户可改
HTML 风格	浅色暖白	可扩深色 / 暗紫调（暂未实现）
Sub-agent 派出方式	一次性 5 个并行	不要分两批

Parameter	Default	Adjustable
Number of Slices	4 (A-D) + 1 (E) = 5	3-6 (E must be included)
Intervention Level	1 time each in Stages 1/2/4	User can authorize skipping Stages 2 and 4 entirely
Deliverable Directory Name	`<Topic>Learning/`	Can be modified by the user
HTML Style	Light warm white	Can be extended to dark / dark purple (not yet implemented)
Sub-agent Dispatch Method	5 dispatched at once in parallel	Do not split into two batches

不适用于（边界划清）

Not Applicable (Clear Boundaries)

"X 行不行 / 为什么 Y / 评估 X 可行性" →
```
long-research
```
"X 领域有哪些好玩案例 / 看看大家在玩什么" →
```
case-radar
```
（散点 vs 体系）
"把这堆素材整理成 HTML" →
```
readable-output
```
（处理已有素材 vs 自己挖素材）
"写文章 / 出初稿" →
```
writing-assistant
```
"做设计稿 / UI 原型" →
```
design-exploration
```
/
```
macos-product-design
```
"PRD / 需求文档" →
```
prd-doc-writer
```

"Is X feasible / Why Y / Evaluate X's feasibility" →
```
long-research
```
"What interesting cases are there in X / See what everyone is experimenting with" →
```
case-radar
```
(scattered vs systematic)
"Organize these materials into HTML" →
```
readable-output
```
(processing existing materials vs independent material gathering)
"Write an article / produce a draft" →
```
writing-assistant
```
"Create design drafts / UI prototypes" →
```
design-exploration
```
/
```
macos-product-design
```
"PRD / requirement documents" →
```
prd-doc-writer
```

参考实例

Reference Example

首发实战：项目根目录

提示词学习/

（2026-05-14 跑通的真实案例）

主题：提示词学习（4 切片全选）
实际耗时：~30 分钟
产物：5 份调研笔记 ~180KB → HTML 学习材料 120KB（1635 行）
卡点教训：见 anti-patterns.md 的"Agent B socket 崩溃"案例

后续每跑一个新主题，建议在

examples-index.md

里记一笔，作为后续主题的参考。

Initial Practical Application: Root directory

提示词学习/

(real case completed on 2026-05-14)

Topic: Prompt Learning (all 4 slices selected)
Actual time taken: ~30 minutes
Deliverables: 5 research notes ~180KB → HTML learning materials 120KB (1635 lines)
Checkpoint lessons: See the "Agent B socket crash" case in anti-patterns.md

After running a new topic, it is recommended to make a note in

examples-index.md

as a reference for future topics.