← 返回 Skills 市场
cellinlab

深度写作

作者 cellinlab · GitHub ↗ · v0.1.0 · MIT-0
cross-platform ✓ 安全检测通过
94
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install cell-deep-writer
功能描述
Adaptive deep-writing workflow for all forms of deep content creation in Chinese, including turning transcripts and note piles into articles, deepening rough...
使用说明 (SKILL.md)

Deep Writer

Overview

Handle deep content writing in three common modes:

  • source-driven: turn transcripts, notes, interviews, or messy drafts into strong articles
  • draft-deepening: take an existing article or outline and make it deeper, clearer, and more complete
  • topic-driven: write a deep piece from a topic, thesis, question, angle, or user idea

Default to pausing for user confirmation after analysis and structure unless the user explicitly asks for one-pass output.

Quick Start

  1. Choose the writing mode before doing anything else.
  2. Infer the audience, scenario, purpose, and stance before drafting.
  3. Complete Stage 1 before writing the article body.
  4. Lock Stage 2 before expanding long-form prose.
  5. Run the final draft against the brief and the quality checklist.

If the user only wants a brief, outline, or structural diagnosis, stop at the relevant stage instead of forcing a full article.

Default Contract

Treat these as the default assumptions unless the user says otherwise:

  • Required: either source material, or a topic / thesis / question / outline to write from
  • Optional: supporting materials, target audience, publication scenario, stance, length, depth, tone, and style
  • Default mode selection:
    • source-driven for transcripts, notes, interviews, speeches, and fragmented material
    • draft-deepening for rough drafts, partial articles, and existing outlines
    • topic-driven for requests that start from a theme, thesis, question, or desired article idea
  • Default language: Chinese
  • Default goal: formal publishable writing, not chat-style rewriting
  • Default structure policy:
    • source-driven: preserve the source's macro frame and allow only micro-adjustments
    • draft-deepening: keep the strongest existing frame, then strengthen weak or thin sections
    • topic-driven: design the best-fit structure around the brief and thesis
  • Default process: wait for confirmation after Stage 1 and Stage 2

If the user provides an already-approved outline, treat it as Stage 2 input and move to Stage 3 after a lightweight brief alignment.

Workflow

Stage 1: Analysis and Planning

Perform all of the following before drafting:

  • identify the writing mode
  • identify the likely audience, scenario, purpose, and stance
  • build a brief using assets/brief-template.md
  • do the relevant analysis for the mode:
    • source-driven: preprocess spoken-language noise, repetitions, and fragmentary clauses
    • draft-deepening: identify weak logic, thin sections, repeated claims, and missing support
    • topic-driven: clarify the central question, working thesis, key angles, and likely counterpoints
  • extract viewpoint units, argument units, or section units and cluster overlapping ideas
  • explain the main logic relations or argument chain

Output at least:

  • the brief
  • mode diagnosis
  • preprocessing notes or argument notes
  • viewpoint clusters or argument clusters
  • logic chain or logic graph summary
  • open questions, evidence gaps, or assumptions when they materially affect the piece
  • updated todo status
  • a confirmation prompt for the next stage

Read references/workflow.md when the material is fragmented, the thesis is still fuzzy, or the structure is not yet obvious.

Stage 2: Structure Design

Use the accepted Stage 1 output to design the writing frame.

Do all of the following:

  • choose the right structure rule for the mode:
    • source-driven: preserve the source's core framework unless the user explicitly requests a rewrite
    • draft-deepening: keep the strongest existing structure, but rebuild weak branches when the draft cannot support the thesis
    • topic-driven: build the strongest structure around the thesis, reader, and scenario
  • merge duplicate sections or overlapping arguments
  • decide section order with the minimum necessary reordering in source-driven mode, or the clearest reader path in topic-driven mode
  • assign each section a clear job, not just a title
  • explain which parts were kept and which parts were adjusted

Use assets/outline-template.md when the user wants a reusable structure blueprint.

Output at least:

  • the optimized structure
  • the role of each section
  • notes on preserved versus adjusted structure
  • updated todo status
  • a confirmation prompt for the next stage

Stage 3: Content Writing

Write the full article only after the structure is locked.

Do all of the following:

  • follow the accepted structure instead of improvising a new one
  • deepen claims with relevant background, mechanisms, examples, or implications
  • remove transcript residue and spoken fillers when they exist
  • keep the author's core meaning intact in source-driven mode
  • keep the improved version aligned with the original draft's intended thesis in draft-deepening mode
  • keep the thesis coherent and non-repetitive in topic-driven mode
  • finish with a quality-check summary

Read references/quality-bar.md before finalizing the draft.

Output Format

Use the staged wrapper in assets/stage-output-template.md unless the user requested another format.

Default wrapper:

【阶段X完成】

---
【本阶段输出】
[stage output]

---
【待办事项状态】
- [completed] 第一阶段:分析与规划
- [pending] 第二阶段:结构设计
- [pending] 第三阶段:内容撰写

---
是否继续下一阶段?(输入“继续”进入下一阶段,或提出修改意见)

For the final stage, replace the confirmation prompt with a short quality-check summary.

Hard Rules

Do not:

  • invent unsupported facts, data, case details, or quotations
  • invent a new central thesis that is absent from the source material in source-driven mode
  • radically rebuild the structure after Stage 2 is accepted
  • keep obvious duplicate viewpoints in separate sections
  • add generic filler that does not deepen the reader's understanding
  • distort the source's stance in order to sound smoother or more authoritative
  • pretend a topic-driven article is well evidenced if the user did not supply evidence and none is available

Always:

  • choose the mode explicitly when the request is ambiguous
  • make the logic chain visible
  • distinguish source viewpoints from supplemental interpretation when that boundary matters
  • keep the brief as the baseline for structure and drafting decisions
  • state assumptions when the user did not provide audience, scenario, or stance

Resource Map

安全使用建议
This skill is instruction-only and internally consistent with its stated purpose. It does not install software or ask for credentials. Before using, avoid pasting highly sensitive secrets or private data into the materials you submit (transcripts, notes, drafts), and optionally verify the upstream GitHub repo if you want provenance or licensing information. If you want stricter control, you can disable autonomous skill invocation on your agent or run the skill only on demand.
功能分析
Type: OpenClaw Skill Name: cell-deep-writer Version: 0.1.0 The skill is a well-structured writing assistant designed to transform transcripts, notes, and drafts into deep Chinese articles. It implements a rigorous three-stage workflow (Analysis, Structure, and Content Writing) with clear quality controls and templates (e.g., SKILL.md, workflow.md, and quality-bar.md). No indicators of data exfiltration, malicious execution, or prompt-injection attacks were found; the logic is entirely focused on content creation and logical consistency.
能力评估
Purpose & Capability
Name/description (deep-writing in Chinese) align with the skill's assets and runtime instructions. The skill only references writing templates, workflow and quality check references included in the bundle; it does not request unrelated binaries, cloud credentials, or system-level access.
Instruction Scope
SKILL.md confines actions to analyzing user-supplied material, building briefs/outlines, and producing staged drafts using local templates and references. It explicitly instructs reading the included assets and references and does not direct the agent to read system files, environment variables, or send data to external endpoints.
Install Mechanism
No install spec and no code files — the skill is instruction-only and will not write or execute external code on install. This is the lowest-risk install posture.
Credentials
No required environment variables, credentials, or config paths. The runtime instructions operate solely on the user-provided source material and the skill's included templates/references.
Persistence & Privilege
always is false and the skill does not request elevated persistent privileges. disable-model-invocation is false (normal), so the agent could call the skill autonomously per platform defaults — this is expected and not excessive in isolation.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install cell-deep-writer
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /cell-deep-writer 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v0.1.0
Initial public release
元数据
Slug cell-deep-writer
版本 0.1.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 1
常见问题

深度写作 是什么?

Adaptive deep-writing workflow for all forms of deep content creation in Chinese, including turning transcripts and note piles into articles, deepening rough... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 94 次。

如何安装 深度写作?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install cell-deep-writer」即可一键安装,无需额外配置。

深度写作 是免费的吗?

是的,深度写作 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

深度写作 支持哪些平台?

深度写作 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 深度写作?

由 cellinlab(@cellinlab)开发并维护,当前版本 v0.1.0。

💬 留言讨论