DeepSeek Harness Plugin Hub

发布与管理完整 Harness Profiles,发现适合你的插件。

探索

插件目录环境预设文档中心动态

社区

发布插件联系我们报告问题

相关链接

Plugin Hub GitHubDeepSeek Harness 官方项目系统状态隐私说明
© 2026 DeepSeek Harness Plugin HubPowered byPaxTech

独立、非官方社区项目,与 DeepSeek 官方无隶属、授权或背书关系。

Gba Mep Docx Toc — DeepSeek Harness 插件(DSH Plugin)
DeepSeek Harness Plugin Hub
ProfilesPlugins分类动态文档登录管理 Profiles
ProfilesPlugins分类动态文档登录
← Plugins
G

gba-mep-docx-toc

Gba Mep Docx Toc

DeepSeek Harness bundle,注册一个 Word 目录技能:仅使用 Python 标准库,将真实的目录字段插入现有的 .docx 文件,支持分级缩进、点引导符、右对齐页码和 _Toc 书签。

插件会安装到这里;不确定时保持 web。

npx -y @deepseek-ai/dsh plugin --profile web add github:gba-mep/docx-toc#c38be1719b88aca02739041b3244aabd19fab174
README兼容性版本

兼容性与来源证明

Gba Mep Docx Toc 以 gba-mep-docx-toc 发布,当前版本为 1.0.0。Plugin Hub 会校验它的 manifest,并保存精确安装来源,便于复现安装结果。

DSH 兼容范围
*
运行环境
any
发布来源
github
Registry 更新时间
2026/10/3

版本

1.0.0stable
2026/10/3

相关插件

正在加载相关插件…

最新版
1.0.0
DSH
*
HMR
重启进程
Tree shaking
未声明可安全裁剪
解包体积
未提供
文件数
未提供
Surface
any
许可证
MIT
发布源
github
GitHub
★ 0
周下载
0
最近提交
2026/10/4
查看源码 ↗
README Badge

点击下方 Badge 复制 Markdown,粘贴到 README 即可。

这是你的 Plugin?认领权益 · 优先安全扫描

验证 package.json 声明的 GitHub 仓库,即可管理这个公开页面。认领后,Hub 会优先安排当前版本的安全扫描,并在通过后公开展示结果。

认领这个 Plugin →
报告问题

相关插件

继续浏览 productivity-workflow 分类下经过校验的插件。

Client Ui Task Board@linxin666/dsh-client-ui-task-board面向 DSH Web GUI 的主机权威任务面板,支持实际会话执行、主机 cron 调度以及可选的跨平台空闲睡眠保护;以挂载方式提供,无需修改 DSH 源代码。Web All@linxin666/dsh-web-allDSH Web UI 全家桶聚合插件:一键安装全部功能插件(task-board / git-graph / pet / remote-web-ui / web-ui-settings / skin-center / community-plugins / compat shim)。compat 桥接层已并入本包(src/client),无需独立 compat npm 包。Deepseek Ipptdeepseek-ipptiPolloWork PPT Studio 及其精选幻灯片模板,作为原生 DeepSeek Harness 对话视图。Codex Ui@michengai/dsh-codex-uiDSH Codex UI — 为 DeepSeek Harness Web 提供 Codex 风格侧栏、工作区会话树、全局搜索和轮次导航

README

gba-mep-docx-toc

A DeepSeek Harness (dsh) bundle that registers one skill: insert a real Word table of contents into an existing .docx.

The skill writes a genuine TOC field with _Toc bookmarks straight into word/document.xml — graded indent per heading level, dot leaders, right-aligned page numbers — using only the Python standard library. It does not need Microsoft Word, LibreOffice, COM, or any third-party package.

Install

dsh plugin --profile web add gba-mep-docx-toc

This repository is also installable straight from source through pnpm's GitHub spec, and it is listed in the awesome-dsh-plugin catalog, so dsh-market can install it with one click.

After installing, restart dsh web once so the bundle layer is composed.

What it does

Given a .docx whose headings use the Heading 1–Heading 9 styles (or carry w:outlineLvl):

  1. finds the headings in document order,
  2. inserts _Toc… bookmarks so every entry is clickable,
  3. builds the TOC \o "1-3" \h \z \u field with one entry per heading — hyperlink anchor → title text → right-aligned tab with dot leader → PAGEREF field,
  4. indents level 2 by 100 and level 3 by 200 character units,
  5. rewrites only word/document.xml and copies every other part of the package byte for byte,
  6. verifies the result with verify (entry count = bookmark count = PAGEREF count).

Requirements

  • A DSH build that exposes the skills service — the entry imports nothing from the harness, so it works on any host publishing that contract (checked against DSH 0.2.0-rc.2).
  • Python 3.9+ on the machine that runs the script. No packages to install.
  • Node.js only because it is a dsh plugin; the entry itself uses node:fs, node:path and node:url.

Usage

python scripts/docx_toc.py build 输入.docx -o 输出.docx --title "目 录" --page-break both
python scripts/docx_toc.py verify 输出.docx
python scripts/docx_toc.py selftest

In a session you can also just describe the task — the skill's description is what the agent routes on.

Boundaries

  • Page numbers are placeholders. Python does not paginate, so real numbers appear only after Word updates the field (Ctrl+A, then F9, choose Update page numbers only). Until then every entry reads 1.
  • No section-based page numbering. With sections that restart numbering (roman front matter, arabic body), the PAGEREF result depends on the Word section setup and is not guaranteed.
  • It does not edit an existing TOC. A document that already carries a TOC field gets a second one; delete the old one first.
  • Paginating is not guaranteed to match Word. Rendering checks used LibreOffice, whose page breaks differ from Microsoft Word's.
  • Headings must use heading styles. Manually bolded and enlarged text is invisible to the script.

Layout

lib/index.js            bundle entry — registers the skill via ctx.skills.register()
cordis.patch.yml        loader patch: one insert row pointing at this package by name
SKILL.md                skill body (frontmatter: name / description / whenToUse)
scripts/docx_toc.py     the implementation (stdlib only)
references/OOXML-TOC.md notes on the OOXML table-of-contents field
tests/                  offline tests: entry behaviour and installability contract
screenshots.json        screenshots shown by plugin storefronts

Tests

node --test tests/

tests/entry.test.mjs runs apply() against a minimal context and asserts the registered fields, the resourceBase, and that the referenced files ship with the package — including a CRLF regression for the frontmatter parser. tests/bundle-contract.test.mjs asserts the installability contract: dsh.bundle declared, patch row naming the package, files whitelist complete, entry importing only node: builtins.

License

MIT — see LICENSE.md.