← Back to Skills Marketplace
eithyhe

book_writer

by EithyHe · GitHub ↗ · v1.0.0
cross-platform ⚠ suspicious
401
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install bookwriter
Description
Generates structured, argument-driven book manuscript sections using modular 800--1000 word conceptual units.
README (SKILL.md)

Book Manuscript Writer

Overview

This skill is designed exclusively for writing book manuscripts, theoretical chapters, and argument-driven academic essays.

It does NOT generate IMRaD-style conference or journal papers.

The fundamental unit of writing is:

A viewpoint-style subtitle
→ followed by a structured 800--1000 word argument unit.

workflow

1. Understanding the main point

When asked to write a book chapter:

  1. Clarify the topic and scope with the user

    • What is the central argument or core claim of this chapter?
    • Who is the intended audience (e.g., general academic readers, specialists, interdisciplinary scholars)?
    • What is the desired length (approximate word count or page range)?
    • Are there specific structural emphases required (e.g., case analysis, theoretical construction, conceptual integration)?
  2. Require a chapter outline from the user

    • The outline must specify the sequence of unit-level claims — not just topic labels. Each entry should express a proposition.

    • Each outline entry should indicate:

      • The core claim of that unit

      • Its role in the chapter's argument arc (what it establishes, challenges, or advances)

      • Key concepts, cases, or sources it will mobilize

    • If the user provides only topic labels (e.g., "Section 3: Social Media"), ask them to convert each into a claim (e.g., "Section 3: Platform Algorithms Reshape Collective Attention Rather Than Merely Reflecting It").

    • If no outline is provided, do NOT proceed to generation. Instead, collaborate with the user to construct one first.

  3. Gather context if needed

  • A user-specified directory containing selected literature or reference documents

  • Supplied research materials, empirical data, or cited sources

  • The relevant theoretical, methodological, and domain background

2. Chapter Structure

A chapter is not a collection of loosely related paragraphs. It is an argument arc composed of discrete, load-bearing units. Each unit advances one identifiable proposition; together, they form a chain of reasoning that moves the chapter from its opening question to its concluding position.

  1. A chapter is a sequence of argument units. Each unit is a self-contained analytical move of 800–1000 words, organized around a single core claim crystallized as a subtitle. The subtitle is not a topic label — it is a compressed thesis. For example, "The Limits of Rational Choice" is a label; "Rational Choice Fails When Preferences Are Endogenous" is a claim.

  2. Units are ordered by logical dependency, not by topic proximity. Unit N must create the conditions — conceptual, evidential, or logical — that make Unit N+1 possible. If two units can be swapped without loss of coherence, the chapter's argumentative spine is weak and must be restructured.

  3. A chapter typically contains 4–7 units. Fewer than 4 suggests the argument is underdeveloped; more than 7 suggests the chapter tries to do too much and should be split.

Every unit must pass a single gatekeeping question: can this unit be removed without weakening the chapter's argument? If yes, the unit fails — it is decoration, not structure. Cut it or reconceive it until it becomes load-bearing.

3. Unit Generator

Each unit FOLLOWS the five-phase structure below. These phases are not optional sections to fill in — they are functional stages of an analytical move. A unit that skips a phase will be structurally incomplete.

Phase 1: Opening Claim (Positioning) — 3–5 sentences

Directly state the unit's core proposition.
The subtitle must express a defensible, non-trivial claim —
not a neutral description or general background.

Required:
-   A clear argumentative position (what this unit asserts)
-   Conceptual direction (where the reasoning will head)

Test: If the opening can belong to a textbook summary,
      it is too neutral. Rewrite.


Phase 2: Tension or Problem Field

Explain WHY the claim matters by surfacing the friction it addresses.
This may take the form of:
-   An empirical case that resists easy explanation
-   A practice dilemma where existing frameworks fall short
-   A conceptual conflict between competing accounts
-   A theoretical ambiguity that prior work has glossed over

Narrative and example are permitted only insofar as they serve analysis.
Any story told here must generate a question, not merely illustrate a point.


Phase 3: Analytical Development (Core Section)

This is the structural center of the unit —
where the actual intellectual work happens.
Must include:
-   Concept clarification: define or sharpen the key terms at stake
-   Logical decomposition: break the claim into its constituent parts
-   Mechanism explanation: show HOW or WHY the claimed relationship holds
-   Structured reasoning: build the argument through explicit inferential steps

Required: At least two explicit logical progression markers
(e.g., however, therefore, further, in contrast, this implies).
Their presence is a proxy for actual argumentative movement —
if the prose flows without them, it is likely describing rather than reasoning.


Phase 4: Conceptual Elevation

Move from the specific to the abstract.
This phase transforms the analytical work of Phase 3
into a broader intellectual contribution:
-   Introduce or refine a concept
-   Reframe the reader's understanding of a familiar phenomenon
-   Shift the interpretive lens through which the problem is viewed

Test: The reader should see the issue differently after this phase
      than they did at the beginning of the unit.
      If perception is unchanged, the elevation has failed.


Phase 5: Closure and Structural Contribution

Conclude by answering three questions explicitly:
1.  What new understanding has emerged from this unit?
2.  How does this advance the chapter's main thread —
    what does the next unit now have access to that it didn't before?
3.  Why is this unit structurally necessary —
    what would collapse without it?

Test: If closure does not produce advancement
      (i.e., the chapter's argument is in the same position
      as before the unit began), regenerate the unit.

Each 800–1000 word unit must satisfy ALL four criteria:

  • Contains at least one identifiable theoretical move (not merely a claim, but a shift in conceptual territory)
  • Includes abstract-level conceptual articulation (not only concrete examples or empirical description)
  • Produces structural advancement (the chapter's argument is measurably further along)
  • Is indispensable to the chapter's argument (passes the removal test)

A unit that meets three of four is a draft. A unit that meets fewer than three should be discarded and reconceived from scratch.

Content Generation Process

Step-by-step approach:

  1. Validate the outline
    • Confirm the user has provided a chapter outline with unit-level claims
    • Verify each outline entry specifies: core claim, role in the argument arc, key concepts/cases/sources
    • Check that the unit sequence follows logical dependency — Unit N must create the conditions for Unit N+1
    • If the outline contains fewer than 4 or more than 7 units, discuss with the user whether to expand or split
  2. Draft units iteratively
    • Generate one unit at a time, following the five-phase structure (Opening Claim → Tension → Analytical Development → Conceptual Elevation → Closure)
    • Start with the pivot unit — the unit that makes the chapter's most consequential theoretical move (typically Unit 3 or 4)
    • Then draft backward — write the units that build toward the pivot
    • Then draft forward — write the units that extend from the pivot toward the chapter's concluding position
    • Write the chapter opening last
    • Run the Unit Quality Gate after each unit before proceeding to the next
  3. Build inter-unit transitions
    • Review the closure of each unit against the opening of the next
    • Phase 5 of Unit N should connect directly to Phase 1 of Unit N+1
    • Add bridging language where the logical handoff is not self-evident
    • Eliminate redundant restatements that slow momentum
  4. Integrate sources and references
    • Embed references within Analytical Development (Phase 3) where they support reasoning
    • Each unit should engage with at least one identifiable source or intellectual tradition
    • Distinguish between sources used as evidence, sources used as interlocutors, and sources used as conceptual tools
    • Compile a complete reference list and verify all in-text citations
  5. Review and refine
    • Argument coherence: read only the subtitles in sequence — they should form a legible argument on their own
    • Indispensability test: for each unit, ask whether the chapter's argument would collapse without it
    • Conceptual consistency: key terms should carry the same meaning throughout, any shift in meaning must be explicitly marked
    • Progression check: the chapter must end in a different intellectual position from where it began
    • Elevation audit: at least half of the units should contain a Phase 4 move that genuinely reframes understanding

DESIGN PHILOSOPHY

Book writing is treated as structured intellectual advancement through modular argument units.

Each section functions as a self-contained yet structurally necessary conceptual movement.

Examples

Example 1: User asks to write a chapter on a broad topic

User: "帮我写一章关于人工智能对教育评价体系影响的内容"

Response:
1. Clarify scope:
   "我来帮你撰写这一章。先确认几个问题:
   - 这一章的核心论点是什么?比如'AI正在瓦解工业时代的标准化评价逻辑'
     还是'AI使多维评价从理念走向技术可行'?
   - 目标读者是教育研究者、政策制定者,还是更广泛的教育工作者?
   - 预期篇幅(5000字、8000字、10000字)?
   - 这一章在全书中的位置——前面的章节已经建立了哪些概念基础?"

2. Require outline:
   "请提供本章的单元级大纲,每个条目需要是一个命题而非标签。例如:
   - ✗ '第一节:传统评价体系'(标签)
   - ✓ '第一节:标准化考试的底层逻辑是工业时代对可量化劳动力的筛选需求'(命题)
   如果你目前只有大致思路,我们可以先一起把大纲搭建出来。"

3. After outline is confirmed, draft units iteratively:
   - Identify the pivot unit (e.g., "AI使过程性数据采集成为可能,
     从而瓦解了终结性评价的垄断地位")
   - Draft pivot unit first, then backward, then forward
   - Run Unit Quality Gate after each unit
   - Build inter-unit transitions
   - Write chapter opening last

Example 2: User provides an outline and asks to generate

User: "我已经有大纲了,请根据以下大纲帮我写第三章:
 1. 知识的本质正在从确定性走向概率性
 2. 课程体系的学科边界在AI时代失去了认识论基础
 3. 动态知识图谱为跨学科课程提供了技术可能
 4. 学习内容的更新速度必须匹配知识生产的速度
 5. 未来课程的核心不是知识本身而是知识生产的方法"

Response:
1. Validate outline:
   - 5个单元,数量合理
   - 每个条目都是命题,符合要求
   - 检查逻辑依赖:单元1(认识论基础)→ 单元2(制度性后果)
     → 单元3(技术条件)→ 单元4(时间维度要求)→ 单元5(范式结论)
   - 逻辑链条成立,确认进入生成

2. Identify pivot unit:
   - 单元3是枢纽——它将前两个单元的理论批判转化为建设性方案,
     为后两个单元提供技术前提

3. Draft in order: 单元3 → 单元2 → 单元1 → 单元4 → 单元5 → 章节开头
   - Each unit follows five-phase structure
     (Opening Claim → Tension → Analytical Development
     → Conceptual Elevation → Closure)
   - Each unit 800-1000 words
   - Run Unit Quality Gate after each

4. Build transitions and integrate references
5. Review: read subtitles in sequence, confirm argument arc holds

Example 3: User provides reference materials and a vague direction

User: "我有一些关于AI教师角色转变的文献资料,放在 /docs/teacher_role/ 目录下,
帮我写一章关于教师角色的内容"

Response:
1. Read reference materials in the specified directory

2. Clarify direction:
   "我已阅读了你提供的文献。在此基础上需要确认:
   - 本章的核心主张是什么?文献中有几种不同取向:
     a) 教师从知识传授者转变为学习设计师
     b) AI替代了教师的部分功能但强化了其不可替代的部分
     c) 教师角色的转变是整个教育系统重构的缩影
     请选择一

Resources

references/

  • writing_style_guide.md: Detailed academic writing conventions extracted from example papers
  • ieee_formatting_specs.md: Complete IEEE formatting specifications
  • acm_formatting_specs.md: Complete ACM formatting specifications

Important Notes

  • Always ask for clarification on topic scope before starting
  • Quality over speed: Take time to structure properly and write clearly
  • Cite appropriately: Academic integrity requires proper attribution
  • Be honest about limitations: Acknowledge gaps or constraints in the research
  • Maintain consistency: Terminology, notation, and style throughout
  • User provides the research content: This skill structures and writes; the user provides the technical contributions and findings
Usage Guidance
This instruction-only skill appears coherent for drafting argument-driven book chapters. Before using it: (1) Only provide literature/files you intend the agent to read — do not upload credentials, private keys, or unrelated sensitive documents. (2) Review any generated citations or quoted passages for accuracy and potential plagiarism — the skill can invent references or misattribute sources. (3) If you enable autonomous invocation for agents, be aware the agent could call the skill without a fresh prompt; restrict autonomous permissions if you don’t want that behavior. (4) If you need the manuscript in a specific language/register, supply clear examples and the required outline as the skill insists it will not proceed without one.
Capability Analysis
Type: OpenClaw Skill Name: bookwriter Version: 1.0.0 The `SKILL.md` file instructs the AI agent to 'Gather context' by reading 'user-specified directory containing selected literature or reference documents.' While this capability is plausible for the skill's stated purpose of academic writing, it introduces a significant vulnerability risk, such as Local File Inclusion (LFI) or information disclosure, if a malicious user could manipulate the input to read arbitrary file paths (e.g., sensitive system files or user credentials). This represents a potential prompt injection attack surface, but the skill's instructions themselves do not demonstrate malicious intent, only a risky capability.
Capability Assessment
Purpose & Capability
The name/description (book manuscript writer) matches the provided SKILL.md and reference material: the skill defines a constrained writing workflow, demands user outlines, and includes style/formatting references. There are no binaries, installs, or environment variables required that would be unrelated to writing.
Instruction Scope
The runtime instructions are detailed and focused on producing argument-driven 800–1000 word units. The SKILL.md does allow the agent to 'gather context' such as a user-specified directory of literature or supplied research materials — this is reasonable for a writing task but means the agent may be asked to read user-provided files. The skill does not instruct reading arbitrary system files or exfiltrating data to external endpoints.
Install Mechanism
No install spec and no code files — instruction-only. This is lowest-risk: nothing will be downloaded or written to disk by an install step.
Credentials
The skill declares no required env vars, credentials, or config paths. The instructions reference user-supplied materials (e.g., a directory of papers) but do not request secrets or unrelated service keys.
Persistence & Privilege
always is false and there is no indication the skill modifies other skills or requests permanent elevated presence. disable-model-invocation is false (the default) so the agent could invoke the skill autonomously; that is normal and not in itself a concern given the otherwise narrow scope.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install bookwriter
  3. After installation, invoke the skill by name or use /bookwriter
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Book-manuscript-writer v1.0.0 — Initial Release - Generates argument-driven book manuscript sections in modular, 800–1000 word units. - Enforces a five-phase analytical structure for each unit: Opening Claim, Tension, Analytical Development, Conceptual Elevation, and Closure. - Requires detailed, claim-based outlines before manuscript generation; collaborates with users to develop these if needed. - Ensures chapters build a rigorous argument arc, with each unit serving an indispensable conceptual role. - Designed exclusively for theoretical book chapters and essays, not for IMRaD-style academic articles.
Metadata
Slug bookwriter
Version 1.0.0
License
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is book_writer?

Generates structured, argument-driven book manuscript sections using modular 800--1000 word conceptual units. It is an AI Agent Skill for Claude Code / OpenClaw, with 401 downloads so far.

How do I install book_writer?

Run "/install bookwriter" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is book_writer free?

Yes, book_writer is completely free (open-source). You can download, install and use it at no cost.

Which platforms does book_writer support?

book_writer is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created book_writer?

It is built and maintained by EithyHe (@eithyhe); the current version is v1.0.0.

💬 Comments