agentsclimarketplace

Lit pdf zotero

Skill qiaomiaojoe/ai-academic-workflow/skills/lit-pdf-zotero

AI学术工作流 · 乔淼PhD「AI学术训练营」skills 库：选题 / 文献搜索 / 文献分析 / 研究设计 / 数据分析 / 全文初稿，30 个 skills 拆到步骤颗粒度，单个可独立运行、可自由组合（Claude Code / Codex）

Install
npx -y skills add qiaomiaojoe/ai-academic-workflow --skill lit-pdf-zotero

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 5 stars5 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

文献 PDF 获取 + Zotero 入库 + 可视化（乔淼PhD · AI学术训练营 · AI学术工作流）。对一份文献清单批量：确认 DOI/CNKI ID → 补全引用数 → 多渠道找 PDF 落盘（OA/preprint/sd-download/wos-download/cnki-download）→ 按金字塔结构建 Zotero Collection 逐条入库（含标签与 extra 元数据）→ 生成金字塔分布图 + PXYV 罗盘投影双图 HTML。Trigger on："把这批文献下载 PDF 入库 Zotero"、"文献入库+可视化"、"batch download PDFs and import to Zotero"、"入 Zotero 并画文献金字塔"。

SKILL.md

4.3 KB, ~1.4k tokens by cl100k_base, as published. Nobody here has run it

Lit PDF + Zotero · 获取 PDF → 入库 → 可视化

对一份文献清单（最好是 lit-pyramid 的分层清单；无分层也能跑，只是入库不分子集合、图 1 跳过）执行四步。

调用约定（独立运行）

  • 输入：文献清单（分层表 / 普通表 / DOI 列表均可）+ Zotero Collection 名（默认 = 项目名或用户给的短名）。
  • 落盘（默认当前工作目录）：PDF 存 pdf_download/，命名 AuthorYear.pdf；最终清单写 文献清单.md；可视化写 [slug]-visualization.html。
  • 工具适配：Zotero MCP 不可用 → 只做 PDF + 清单落盘，并告知手动导入方式；下载工具按环境可用性挑。
  • 调用方覆盖约定：调用方（工作台 prompt / 用户）显式给出的参数、输入路径、落盘路径与工具选择，一律覆盖本 skill 的默认；但本 skill 的方法步骤、硬约束与停点不可被省略或稀释——调用方若要求跳过某硬约束，以本 skill 为准并提示冲突。

Step 1 · 确认 DOI / URL / CNKI ID

英文论文 → DOI；中文论文 → CNKI ID + 期刊+期号；经典专著 → 出版商 URL 或 DOI；非学术来源 → URL。

Step 2 · 补全引用数

英文用 OpenAlex API（https://api.openalex.org/works/doi:{DOI} 取 cited_by_count）或 Google Scholar；中文用 cnki-paper-detail。清单里已有的不重采。

Step 3 · 查找 PDF 并落盘

  • 英文优先 sd-download / wos-download → Open Access（Unpaywall / OpenAlex / 出版商 OA）→ 作者自存档 → preprint（arXiv / SSRN / OSF）
  • 中文用 cnki-download（要求已登录）
  • 找不到的标 "no-pdf"；不从可疑来源硬凑。

最终清单落盘（下游文献分析在 Zotero 不可用时的兜底元数据）： | # | 标题 | 作者+年份 | 金字塔层 | PXYV 节点 | DOI/CNKI ID | 被引数 | 核心论点 | PDF 状态 |（状态 = open-access / preprint / author-copy / cnki / no-pdf）

Step 4 · 入库 Zotero + 可视化

Collection 结构：顶层 = Collection 名，下设 Classic / Key Texts / Supporting-P / Supporting-X / Supporting-Y / Supporting-V（无分层信息则只建顶层）。

逐条入库：有 DOI → zotero_add_by_doi(doi, collections=[对应子集], tags=[金字塔层, PXYV节点, 项目slug], attach_mode="auto")；只有 URL/CNKI → zotero_add_by_url（或用 gs-export / sd-export / wos-export / cnki-export 直接推送）。入库后用 zotero_update_item 写 extra 字段：

citations: 123
core_argument: 一句话核心论点
pxyv_weights: P=0.7, V=0.3

（pxyv_weights 按"主导维度"生成，单一维度 = 1.0。标签：金字塔层 "classic"/"key"/"supporting"；节点 "P"/"X"/"Y"/"V" 可多选；项目 slug。）

入库报告：| # | 标题 | 入库状态 | PDF 附件 | Collection | 标签 | + 统计（总入库 / PDF 已附 / 金字塔分布 / 节点覆盖）。

可视化（单文件 HTML，纯 HTML+CSS+内联 SVG 无外部库；浅色背景 #FAF7F2，amber #B8834A）：

  • 图 1 文献金字塔分布图：三层水平堆叠，顶 Classic（窄，标篇名+作者）→ 中 Key（标标题+主要节点）→ 底 Supporting（按数量宽度）；Classic #E65100 / Key #1565C0 / Supporting #558B2F。
  • 图 2 PXYV 罗盘投影：800×800 SVG；四锚点 P=(0,+1) / X=(+1,0) / Y=(0,-1) / V=(-1,0)，边缘标大字母 + 颜色（P #B8834A / X #5B8DBE / Y #7BAE7F / V #9B6DB0）。每篇一个圆点：位置 = pxyv_weights 归一化加权质心；半径 = 4 + 6*log10(citations+1) clip [4,24]px；填充 = 金字塔层色 opacity 0.7；描边 = 主导维度色；悬停 tooltip（标题/作者年/引用数/核心论点）；同心圆 0.25/0.5/0.75/1.0 网格 + 双图例。

Gives 0 of the 12 instructions most pdf office docs skills give in ~1.4k tokens

Counted across 635 of the 690 authors here whose files we hold, read 2026-08-06

  • extract text using pdfplumberin 92 of 635, across 25 files
  • create PDFs using reportlabin 83 of 635, across 16 files
  • read FORMS.md to fill out PDF formsin 80 of 635, across 13 files
  • OCR scanned PDFs using pytesseractin 77 of 635, across 10 files
  • merge or split PDFs using qpdfin 70 of 635, across 3 files
  • use Excel formulas instead of hardcoded calculated valuesin 68 of 635, across 12 files
  • unpack edit xml and repack existing documentsin 63 of 635, across 8 files
  • document sources for hardcoded valuesin 61 of 635, across 9 files
  • write minimal python code without unnecessary commentsin 59 of 635, across 7 files
  • run the recalculation script after adding or modifying formulasin 58 of 635, across 6 files
  • fix all identified formula errors and recalculatein 58 of 635, across 6 files
  • format years as text stringsin 57 of 635, across 5 files

Said here and by no other author read

  • confirm DOI or CNKI ID per item
  • complete citation counts via API
  • download PDFs to local directory
  • skip items with existing citation counts
  • find PDFs through multiple channels
  • mark unfound items as no-pdf

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.