dsh-book2skill DSH plugin

dsh-book2skill introduces a 5-stage long-running task workflow to DeepSeek Harness: fetch book, parse chapters, deep read, generate SKILL.md, and install. It includes 3 human gates for review and approval. Tasks are stored in host storage, cross-session resumable, and cancellable at any time.

skills-workflows browser-web

How to install the dsh-book2skill DSH plugin

dsh plugin --profile web add file:/absolute/path/dsh-book2skill-0.1.2.tgz

Copying does not run this command. Review the repository and version before installing the dsh-book2skill DSH plugin.

dsh-book2skill DSH plugin data source

dsh-book2skill DSH plugin snapshot date: Aug 16, 2026

discovered

What the dsh-book2skill DSH plugin can do

EPUB parsing with automatic chapter splitting (merge <200 words, split >5000 words)
PDF parsing with OCR fallback (three-level detection: paddleOcr service, loopback RPC, HTTP endpoint)
Browser timeline panel for 5-stage progress and human gate approval
Cross-session resumable jobs with cancellation support
Integration with z-lib for book downloading (uses existing cookies, no hardcoded keys)

Where the dsh-book2skill DSH plugin fits

Convert a downloaded EPUB book into a structured skill with human quality gates
OCR-process a scanned PDF book and generate a SKILL.md
Review and edit the auto-generated SKILL.md before installation
Resume a book conversion job after server restart

Who the dsh-book2skill DSH plugin is for

DSH users who want to turn books into reusable skills
Users who prefer a guided, multi-stage workflow with human approval steps

dsh-book2skill DSH plugin limitations

Requires DSH >=0.1.0-rc.3 <0.2.0 and Node.js ^22.19.0 || >=24.0.0
OCR depends on external services (dsh-paddle-ocr or custom endpoint) and may fail if none available
Plugin is published as restricted (npm publishConfig.access: restricted)
Browser panel changes require manual rebuild or HMR to reload

dsh-book2skill DSH plugin: from the repository README

Quoted from the omdsh-dev/dsh-book2skill README, the upstream source of the dsh-book2skill DSH plugin. Copyright remains with the original authors.

书籍转技能（Book → Skill）：DSH 插件包，把一个 5 阶段长任务工作流带进 DeepSeek Harness—— **获取书籍 → 解析分章 → 深度阅读 → 生成 SKILL.md → 安装**，中途有 **3 个人类门控**。确定性步骤（EPUB 解析 / PDF 提取 / OCR / 安装复制）是宿主工具；理解与生成（浅读建地图、设计方向问题、深读核心章、撰写 SKILL.md 与自检）由 agent 执行；浏览器面板负责 5 阶段时间线展示与门控审批。任务状态存宿主存储域，**跨会话、跨重启可恢复，可随时取消**。 ## 架构 | 平面 | 内容 | |------|------| | 宿主（`src/index.ts`） | `book2skill` 存储域（jobs 表）、`/book2skill` HTTP 面板路由、10 个 `book2skill_*` 工具 | | 浏览器（`src/client/`） | `conversation.view` 时间线面板（React），纯 fetch 轮询宿主路由 | | 脚本（`scripts/`） | `parse_epub.py`（修复版：标题行不再丢失、>5000 字章节按 `##` 拆分）、`pdf_parse.py`（PyPDF2 探测/分块） | ## 5 阶段 + 3 门控 1. **获取书籍**：本地路径（目录选择器，browse 能力）/ z-lib 搜索下载（复用 book-downloader 的 cookies，无硬编码密钥） 2. **解析分章**：EPUB → 按章 md + toc.md（<200 字合并、>5000 字拆分）；PDF → PyPDF2 先探测（<5000 有效字或乱码判扫描型）→ 扫描型走 OCR，面板显示逐页进度（37/420）与「排队中」 3. **深度阅读【门控1】**：agent 浅读建地图 → 发布 ≤3 个选择题（选项带书中章节背景）→ 用户在面板作答 → agent 深读 3-5 个核心章 4. **生成 SKILL.md【门控2】**：可编辑预览 + 知识地图 + 3 项自检清单（SOP 可溯源 / 索引准确 / 触发词宽窄，不通过标红，可一键让 agent 修）→ [重新生成] / [通过并继续] 5. **安装【门控3】**：目标多选（~/.claude/skills、~/.codex/skills、~/kk_skill/skills 同步仓库）→ 确认后 agent 调 `book2skill_install` → 完成卡（触发示例 chips + 复制即试） ## OCR 解耦（三级探测，无硬依赖）本插件**不 inject** `paddleOcr`（缺失服务会永远挂起本行）。运行时按级探测：

Read the full READMERepository license: BSD-3-Clause

dsh-book2skill DSH plugin questions

How do I install dsh-book2skill?

First run `npm pack` to create a .tgz file, then use `dsh plugin --profile web add file:/path/to/dsh-book2skill-0.1.2.tgz`. After that, restart the dsh web process for host-side changes to take effect. The browser panel will automatically reload if HMR is enabled.

What file formats are supported?

The plugin supports EPUB and PDF. For PDF, it attempts to detect if it's a scanned document using PyPDF2. If it is, OCR is triggered. OCR requires an external service like dsh-paddle-ocr or a custom endpoint.

Can I resume a conversion job if the server restarts?

Yes, jobs are stored in host storage and are cross-session resumable. You can also cancel a job at any time using the `book2skill_cancel` tool.

How do I configure the OCR endpoint?

You can set the environment variable `BOOK2SKILL_OCR_ENDPOINT` to point to your OCR service. The default is `http://127.0.0.1:8011/api/pdf/ocr`. Alternatively, you can install the dsh-paddle-ocr plugin which provides the OCR service automatically.

What are the human gates for?

The workflow has three human gates: one after deep reading (where you answer questions to guide the agent), one for reviewing the generated SKILL.md (with self-check items), and one for confirming the installation target directories. These ensure quality control.