DeepRead — Research

DeepRead is an evidence-first reading analyst that reconstructs an author's claims, evidence, and reasoning from supplied documents, and builds knowledge maps or guides Feynman-style learning. It produces structured analyses such as argument trees, evidence ledgers, and recall questions, tailored to quick, deep, map, Feynman, or whole-book reading modes.

research claude-code

📂 View skill on GitHub ↗ 🚀 How to use

✨ What it does

Verifies source extraction and flags incomplete material before analysis.
States the author's central claim as a proposition, not a topic label.
Builds an argument tree with typed relationships (supports, explains, qualifies, contradicts, illustrates).
Creates an evidence ledger with confidence labels (author's stated position, source fact/data, reasoned inference, unver
Tests reasoning for logical fallacies like correlation-as-causation and overgeneralization.
Synthesizes at the correct scale, including chapter-to-thesis links for whole books.

🎯 When to use it

When the user asks for a deep reading, close reading, or whole-book understanding of a supplied document.
When the user wants claims separated from evidence, examples, assumptions, and inference.
When the user wants a knowledge map or mind-map-ready hierarchy.
When the user asks to use the Feynman technique or create recall questions.
When the request includes Chinese triggers like 精读, 核心观点, 论证逻辑, 知识地图, 思维导图, 费曼读书法, or 整本书.

🚀 How to use

Trigger DeepRead by asking for a deep reading, close reading, knowledge map, Feynman explanation, or whole-book analysis of a supplied document. Provide the text, file, PDF, or document set. The skill defaults to deep mode unless you specify quick, map, feynman, or book. Example prompts:

精读这篇文章，提取核心观点和论证逻辑。

Use the Feynman technique to help me learn this chapter.

📄 Output: A structured analysis report including source status, synthesis, central claim, argument tree, evidence ledger, concepts, assumptions, and recall questions (mode-dependent).

📦 Add this skill to Claude Code

# 1. Get the skills repo
git clone --depth 1 https://github.com/alirezarezvani/claude-skills /tmp/claude-skills

# 2. Copy this skill into your project (or ~/.claude/skills for all projects)
mkdir -p .claude/skills
cp -r /tmp/claude-skills/research/deepread .claude/skills/deepread

Skill source: research/deepread/SKILL.md

⚠️ Good to know

Requires supplied reading material; does not discover sources (use deep-research) and is not for conventional executive summaries (use research-summarizer).

❓ FAQ

What is the difference between quick and deep mode?

Quick mode gives the gist: thesis, up to three supporting claims, key evidence, and three questions. Deep mode provides a full argument tree, evidence ledger, concepts, assumptions, gaps, and counterarguments.

Can I combine modes?

Yes, combine modes when you need both comprehension and retention, for example book followed by feynman.

How does DeepRead handle long documents?

For material longer than roughly 9,000 words, it splits on semantic boundaries, analyzes each part, then runs a separate synthesis pass.

🤖 Overview, features, install steps and FAQ were generated from the project's SKILL.md on Sep 4, 2026. Always check the original source before running commands.