🧭 Web Novel Tutorial Curation Skill

💡 A reusable AI-agent skill for extracting, summarizing, classifying, validating, and integrating web-novel / fiction-writing tutorial materials and excellent-novel case libraries into clean knowledge libraries.

💡 This repository contains a portable SKILL.md package designed for AI agents that work with Chinese web-novel, online fiction, genre fiction, and long-form serial-writing resources. The workflow is also useful for any agent that must process tutorial websites, writing-course pages, local documents, EPUB/PDF/DOCX resources, or folders of benchmark fiction cases without copying raw copyrighted material.

🎯 What This Skill Solves

💡 AI agents often make avoidable mistakes when processing tutorial sources:

• ✅ scraping only visible page text while missing APIs or server-rendered data;
• ✅ assuming all article detail pages share one template;
• ✅ copying full tutorial bodies instead of producing reusable summaries;
• ✅ treating finished novels as tutorials and producing plot reports instead of reusable craft analysis;
• ✅ using broad keyword tags that classify almost every article incorrectly;
• ✅ updating master prompts or skills before validating counts and files;
• ✅ leaving project-specific details inside generic agent instructions;
• ✅ forgetting page ranges, filters, extraction dates, and limitations.

💡 This skill turns those lessons into a repeatable workflow.

🎯 Core Capabilities

• ✅ Discover whether tutorial data comes from HTML, rendered DOM, APIs, script state, documents, or mixed templates.
• ✅ Extract list pages before detail pages.
• ✅ Deduplicate by stable IDs or canonical URLs.
• ✅ Detect low-text, image-only, video-only, blocked, or changed-template pages.
• ✅ Summarize tutorials into writing principles, not copied articles.
• ✅ Analyze excellent-novel case libraries for reader promise, conflict systems, structure, character pressure, suspense, style, and ending payoff.
• ✅ Build Markdown and JSON knowledge libraries.
• ✅ Classify materials by source category and writing task.
• ✅ Validate counts, JSON parsing, encodings, links, tags, and master-file integration.
• ✅ Update prompts, project guides, or existing skills only after validation.

🗂️ Repository Structure

.
├─ README.md
├─ README.zh-CN.md
├─ LICENSE
├─ assets/
│  └─ logo.png
└─ web-novel-tutorial-curation/
   ├─ SKILL.md
   ├─ agents/
   │  └─ openai.yaml
   └─ references/
      ├─ extraction-workflow.md
      ├─ output-templates.md
      ├─ case-library-analysis.md
      ├─ excellent-novel-case-corpus-patterns.md
      └─ quality-gates.md

💡 The actual skill package is the web-novel-tutorial-curation/ folder.

🚀 When To Use

💡 Use this skill when an AI agent is asked to:

• ✅ extract writing tutorials from a website;
• ✅ summarize fiction-writing courses into a reusable knowledge base;
• ✅ process the first N pages of a tutorial list;
• ✅ classify author interviews, platform lessons, or genre guides;
• ✅ summarize a folder of excellent novels or EPUB fiction cases into transferable craft lessons;
• ✅ build a local folder of Markdown and JSON summaries;
• ✅ update master AI prompts or existing writing skills with a new source;
• ✅ validate that extraction results are complete and not stale.

📌 Typical Output

💡 The skill recommends producing at least:

TutorialLibrary/
├─ 00_Overview.md
├─ 01_All_Items.md
├─ 02_Category_Summary.md
├─ 03_Theme_Index.md
└─ tutorial_index.json

💡 Each item should preserve:

• ✅ title;
• ✅ source link;
• ✅ category or column;
• ✅ extraction date;
• ✅ page/rank/order when relevant;
• ✅ available counters;
• ✅ content status;
• ✅ tags;
• ✅ summary;
• ✅ usable experience;
• ✅ limitations.

🧩 Copyright-Friendly Design

💡 This skill is intentionally designed to avoid mirroring tutorial websites.

💡 Recommended output stores:

• ✅ summaries;
• ✅ tags;
• ✅ source links;
• ✅ short headings when useful;
• ✅ operational writing advice.

💡 It does not recommend storing full original article bodies by default. 💡 For finished-fiction case libraries, it also avoids storing full chapters, scenes, or distinctive prose imitation.

📌 Novel Case Libraries

💡 When the source is a folder of novels or benchmark fiction cases, the skill treats it as case evidence rather than tutorial content. It recommends extracting metadata first, then summarizing craft mechanisms such as opening promises, protagonist pressure, conflict systems, information control, motif use, escalation rhythm, prose texture, and ending payoff.

💡 See case-library-analysis.md for the dedicated workflow.

🚀 Installation

💡 Copy the skill folder into your AI agent's skill directory:

web-novel-tutorial-curation/

💡 For Codex-style skills, the important file is:

web-novel-tutorial-curation/SKILL.md

💬 Example Prompt

Use $web-novel-tutorial-curation to extract the first 20 pages of this fiction-writing tutorial list, summarize all articles into Markdown and JSON, validate the counts, and update my master writing-agent prompt with the new source.

Use $web-novel-tutorial-curation to summarize this folder of excellent novel EPUBs as a case library. Build a metadata index, classify each work by craft value, extract transferable writing lessons, note imitation risks, and update my writing-agent source registry.

✅ Quality Gates

💡 Before finalizing, the agent should verify:

• ✅ exact source URL, page range, filter, and extraction date;
• ✅ list count and duplicate handling;
• ✅ detail-page success count and low-text count;
• ✅ JSON parse success;
• ✅ Markdown readability;
• ✅ source links;
• ✅ tag quality;
• ✅ absence of full copyrighted article bodies;
• ✅ master-file updates;
• ✅ skill validation when a skill was edited.

💡 See quality-gates.md for the full checklist.

🌐 Chinese README

🌐 Chinese documentation is available in README.zh-CN.md.

📜 License

💡 MIT License.