AI Agent Hub
Back to skills
Deep Reader icon

Deep Reader

Knowledge Management Updated 2026.08.30

Paste the following prompt into your AI chat to install this skill:

Please follow https://skillhub.cn/install/skillhub.md to install @user_bdacebb4/read-helper

About this skill

Problem Solved

AI summaries can compress a book into conclusions, but readers still need to know why the author wrote it, where the argument is fragile, and which chapters deserve slow reading. This skill is not positioned as a tool that finishes the book for you. It moves mechanical work forward—structure, context, and concept location—so the reader can engage with the text through their own experience. It fits ebook files or title-only reading consultations, with short, scannable output for IM-style conversations.

How It Works

The skill enters different flows based on input state. If the user uploads an EPUB / PDF and expresses reading intent, it can read the full text. With only a title, it answers from model knowledge and may use WebSearch for supplementary context. When asked “is it worth reading,” it switches to a quick-view response. Internally, it classifies books by method type and subject domain, then selects spoiler controls, argument analysis, or practical frameworks, without exposing those labels. Key capabilities include:
- Pre-reading prep: type, author context, book trajectory, core problem, and reading advice.
- Chapter discussion: terms, claims, reasoning chains, and hidden premises.
- Debate mode: Socratic questioning, counter-argument, and thought experiments to test ideas.
- Extended reading: same-direction, opposing, and intellectual-lineage recommendations.
- Auxiliary output: when requested, generate a Markdown mind map and convert it to interactive HTML.

Boundaries

The skill avoids full book reviews or substitute summaries. For narrative works it maintains spoiler controls unless the user explicitly allows spoilers. If the user uploads only a file without a text instruction, it does not start analysis. Extended recommendations may rely on web search; mind maps are generated only after an explicit request.

Use Cases

  • Before reading a political philosophy book, map its author context, argument flow, and chapters worth slow reading.
  • While reading chapter 3, challenge the author's premises, claims, and hidden assumptions in a debate-style review.
  • Before spending ten hours on a sociology book, quickly judge its type, core claims, and whether it is worth reading.
  • After finishing a book, find supporting, opposing, and intellectually related works on the same theme.

Best For

  • Engineers reading nonfiction who want to trace argument chains and concept boundaries in complex books.
  • Product or operations people preparing research and reviews who need to turn book claims into debatable questions.
  • Self-directed interdisciplinary readers who want novel or historical context without spoilers.
  • Readers preparing notes who want a book map and related recommendations around one text.