mcpField
B
book-reader-conversations

book-reader-conversations

github.com/lvtd-llcCreative & media163 installsUpdated 7w ago
Repository

About book-reader-conversations

book-reader-conversations is a skill for authors of useful nonfiction who want to learn from readers before or during drafting. It guides listening and teaching conversations, then turns evidence into book decisions and a next test.

Key features

  1. 01Define a focused learning goal
  2. 02Recruit readers matching a hypothesis
  3. 03Run listening or teaching conversations
  4. 04Synthesize feedback into book decisions
  5. 05Choose a next test or manuscript action

Use cases

  1. 01Validate a nonfiction book promise
  2. 02Identify the reader profile and use case
  3. 03Test whether a table of contents teaches clearly
  4. 04Collect reader language, objections, and examples

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/lvtd-llc/skills --skill book-reader-conversations

After installing, type/book-reader-conversationsin Claude Code to confirm it loaded.

Bundled files

References and scripts installed along with SKILL.md.

agents/
openai.yaml
references/core/
checklist.md
examples.md
knowledge.md
rules.md
workflows/
run-reader-conversations.md

Details

From the project README

Book Reader Conversations

Core Lens

Reader conversations are product discovery for useful nonfiction. The goal is not to ask whether people like a book idea. The goal is to understand the reader's real situation, language, failed attempts, urgency, objections, and ability to receive the value the book intends to deliver.

Use this skill before or during drafting to:

  • Validate a book promise before writing too much prose.
  • Escape the author's curse of knowledge.
  • Find the strongest reader profile and use case.
  • Test a takeaway-first TOC by teaching from it.
  • Turn conversations into scope, structure, examples, and beta-reader leads.

Reference Routing

NeedRead
Core model and terminologyreferences/core/knowledge.md
Interview and synthesis rulesreferences/core/rules.md
Example prompts and synthesis patternsreferences/core/examples.md
Fast setup and review checklistreferences/core/checklist.md
Run a conversation roundworkflows/run-reader-conversations.md

Workflow

1. Define The Learning Goal

Choose one primary question for this round:

  • Is the reader problem real and urgent?
  • Which reader profile cares most?
  • What does the reader already know or believe?
  • Which promise is easiest to understand and recommend?
  • Does the TOC sequence help the reader receive value?

Do not mix early listening with pitching unless the user is explicitly running a teaching conversation.

2. Recruit Readers Who Match The Hypothesis

Start with friendly first contacts, existing readers, community members, customers, students, or people already trying to solve the problem.

Avoid cold outreach unless no warm path exists. If nobody relevant will talk, treat that as a scope or access problem rather than a scheduling problem.

3. Run The Right Conversation Type

Use two modes:

  • Listening conversation: explore lived experience, language, urgency, alternatives, and obstacles.
  • Teaching conversation: help the reader through one slice of the TOC and watch where they get value, confusion, skepticism, or missing steps.
4. Synthesize Into Book Decisions

After each batch, update:

  • Reader profile and exclusions.
  • Promise wording and reader language.
  • TOC order, section titles, and missing steps.
  • Proof, examples, objections, and caveats.
  • Potential beta readers, testimonials, or launch allies.
5. Decide The Next Test

Recommend one of:

  • More listening with a sharper reader segment.
  • Teaching a high-value chapter or section.
  • Revising the TOC with book-toc-lab.
  • Drafting enough for reader-experience-edit or beta-reader-feedback.

Output Format

When helping with reader conversations, return:

  1. Learning goal and target reader hypothesis.
  2. Recruiting list or outreach plan.
  3. Conversation guide with prompts.
  4. Note-taking schema.
  5. Synthesis table: signal, evidence, implication, book decision.
  6. Next test or manuscript action.

Quality Bar

Keep the conversation about the reader's life before the author's idea. Prefer observed behavior, concrete stories, and exact reader language over opinions, compliments, or predictions.

Comments

Markdown supported