dsh-deepread — Claude Code plugin: analysis and setup
The dsh-deepread plugin reads books and articles in five modes, from a quick overview to a knowledge map or Feynman study plan. Use it for research and fact-checking.
MIT ★ 3,785 updated 2026-10-25
/plugin marketplace add davepoon/buildwithclaude
/plugin install dsh-deepread@buildwithclaudeWhat dsh-deepread does
dsh-deepread accepts .txt, .md, .markdown, .html, and .pdf files, web links, or pasted text. Its five modes cover quick summaries, a detailed article read, knowledge mapping, Feynman-style learning, and whole-book analysis. The deep mode is the default.
Reports connect claims with evidence, data, and labeled relations, then assign one of four confidence levels. The plugin can compare sources in a batch, run a budget preflight, and expose progress for background jobs. Results appear as Markdown; requested exports can be written as .md, .mm for FreeMind or XMind, and .html in deepread-output/.
Anti-bot pages require pasted text instead of direct fetching. The plugin runs with Claude Code and Cursor, while the marketplace installation uses Claude Code commands.
What it helps with
- get the point of an article quickly
- trace claims back to supporting evidence
- build a knowledge map before citing a source
- turn reading into a spaced-review plan
- assemble a report from a long book
Example prompts for the agent
- Deep-read
docs/architecture.mdindeepmode and show each claim with its evidence. - Analyze this text in
mapmode, including the data table and recall questions. - Read this book with the
feynmanmode and create a review schedule.
How to install dsh-deepread
Plugin marketplace
/plugin marketplace add davepoon/buildwithclaude/plugin install dsh-deepread@buildwithclaudeStep by step, with commands for each platform: Claude Code plugins: marketplaces, installing, and building your own
Questions about dsh-deepread
How do I install dsh-deepread in Claude Code?
First run /plugin marketplace add davepoon/buildwithclaude, then run /plugin install dsh-deepread@buildwithclaude.
What does dsh-deepread do?
It analyzes books and articles, extracts claims, argument structure, and evidence, then produces structured reports or knowledge maps.
What sources and file types can dsh-deepread read?
You can provide .txt, .md, .markdown, .html, or .pdf files, a web link, or pasted text. Pages protected by anti-bot checks need pasted content.
Source text
dsh-deepread · MIT
# 📖 dsh-deepread Deep-reading assistant for Claude Code: extract core claims, argument structure, and key evidence from books and articles into a structured "claim–evidence–data–relation" report. ## Five modes | Mode | Best for | Key output | | --- | --- | --- | | `quick` | "What is this article about?" at a glance | Summary, core claim, up to 3 arguments, quotes, questions | | `deep` (default) | Reading one article carefully | Core claim, argument structure (claim + evidence), argument flow, key concepts, critical thinking | | `map` | Research and fact-checking before citing | Knowledge map: ten content categories, claim–evidence pairing, key data table, eight relation labels, four confidence levels, Mermaid mindmap, XMind outline, active-recall questions | | `feynman` | Truly learning and teaching it to others | 11-step loop with spaced repetition on days 1/3/7/14/30 | | `book` | Whole books / very long texts | Table of contents, chapter flow, full-book summary assembled from per-part deep reads | One-line picker: in a hurry, `quick`; read one article thoroughly, `deep`; cite and fact-check, `map`; learn and remember, `feynman`; a whole book, `book`. ## Installation ```bash # From the BuildWithClaude marketplace /plugin install dsh-deepread@buildwithclaude # Or from the upstream repository claude plugin install xiehuan123/dsh-deepread npx skills@latest add xiehuan123/dsh-deepread ``` ## Usage Trigger by saying something containing "deep-read / analyze / knowledge map / Feynman", for example: - `Deep-read docs/architecture.md` - `Analyze this article in knowledge-map mode: <paste text>` - `Read this book with the Feynman technique and give me a review plan` - `Quickly summarize this article: https://mp.weixin.qq.com/s/xxxx` Input can be a file path (`.txt/.md/.markdown/.html/.pdf`), a web link (WeChat articles are fetched directly; for anti-bot sites, paste the text), or pasted text. ## Output The report is shown in the conversation as Markdown by default. On request, export to `.md`, `.mm` (FreeMind mindmap, importable by XMind), or `.html` (self-contained web report with light/dark theme) — written to `deepread-output/` in the workspace. ## License MIT
© xiehuan123 · MIT · davepoon/buildwithclaude