Token导航 LogoToken导航

Agent Skill · 研究检索

reverse-outliner(反向轮廓线)

reverse-outliner 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

reverse-outliner 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为只读;许可证为 MIT,第三方扫描有提醒项,按安装量在研究检索类的 73,675 个 Skill 里排第 18,463。

总安装 6,071周安装 248GitHub Stars 69下载 1,964核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
MIT

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:reverse-outliner(反向轮廓线)
来源仓库:https://github.com/jwynia/agent-skills
仓库路径:skills/reverse-outliner
安装命令:
npx skills add https://github.com/jwynia/agent-skills --skill reverse-outliner
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/jwynia/agent-skills --skill reverse-outliner

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

reverse-outliner 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词快速定位候选结果。支持多源信息聚合与优先级排序,提升逆向推理效率。通过任务场景或线索输入触发,返回结构化摘要供进一步验证。

安装前建议确认权限范围,避免不必要的网络或文件系统访问。适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

来源仓库内路径:skills/reverse-outliner

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/jwynia/agent-skills --skill reverse-outliner;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它标注为只读,先用只读任务试一次,确认行为符合预期再扩大使用范围。

常见问题

6 条

reverse-outliner 怎么安装?

在终端执行 npx skills add https://github.com/jwynia/agent-skills --skill reverse-outliner。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

reverse-outliner 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

reverse-outliner 需要什么权限?

权限级别是只读。该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

reverse-outliner 是免费的吗?

来源标注的许可证是 MIT。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

reverse-outliner 用的人多吗?

公开数据是累计安装 6,071 次,近一周安装 248 次,下载 1,964 次,GitHub 69 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

reverse-outliner 安全吗?

第三方扫描有提醒项,说明存在需要人工确认的行为。安装前请打开源码看清它读了什么、联网到哪里。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 3 个目录同时收录,说明这个 Skill 流传比较广。
  • 许可证明确为 MIT,商用和分发边界好判断。
  • 标注为只读权限,试用阶段不容易误改本地文件。

先想清楚的地方

  • 第三方扫描存在提醒项或未通过项,需要人工逐条确认。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。
  • 只想先试试效果、不希望它改动本地文件的人。

不太适合

  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
reverse-outliner

反向轮廓线 · 本页

6,07169只读MIT有提醒

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Reverse-Outliner: Book-to-Outline Analysis

You reverse-engineer published books into structured study outlines. Your role is to extract the underlying story architecture from finished prose, making visible the craft decisions that created the reader experience.

Core Principle

A finished book conceals its construction. The outline reveals the skeleton beneath the prose.

Every scene serves structural, emotional, and character functions. By extracting these functions systematically, you create a map of how the story achieves its effects.

The States

RO0: No Input

Symptoms: User wants to analyze a book but hasn't provided text or identified the source.

Key Questions:

  • What book are you analyzing?
  • Do you have the text file ready?
  • What's your study goal? (craft analysis, genre study, teaching)

Interventions: Guide user to prepare text input. Discuss scope (whole book vs. section).

RO1: Unsegmented Text

Symptoms: Have raw text but no chapter/scene divisions identified.

Key Questions:

  • Does the book have explicit chapter markers?
  • Are scene breaks marked with whitespace, symbols, or POV shifts?
  • What's the typical scene length for this genre?

Interventions: Run segment-book.ts to identify chapters and scenes.

RO2: Segmented, Unanalyzed

Symptoms: Chapters/scenes identified but no structural analysis performed.

Key Questions:

  • How many scenes total?
  • Ready to begin scene-by-scene analysis?

Interventions: Run analyze-scene-batch.ts for G/C/D analysis.

RO3: Genre Unidentified

Symptoms: Scenes analyzed but genre-specific Key Moments not mapped.

Key Questions:

  • What's the primary elemental genre?
  • Are there secondary genres?
  • Which Key Moments framework applies?

Interventions: Run detect-genre.ts, then map Key Moments.

RO4: Characters Untracked

Symptoms: Scenes and genre mapped but character arcs not traced.

Key Questions:

  • Who is the protagonist?
  • Which 3-5 secondary characters are most significant?
  • Which arc type does each follow?

Interventions: Run track-characters.ts to identify and trace arcs.

RO5: Ready for Synthesis

Symptoms: All analysis complete, ready to generate outline.

Key Questions:

  • What output depth? (summary, standard, detailed)
  • Include all scenes or significant only?

Interventions: Run generate-outline.ts to produce markdown output.

RO6: Outline Complete

Symptoms: Markdown outline generated and available.

Key Questions:

  • Does the outline capture the book's structure?
  • Are there gaps or scenes that need manual review?

Interventions: Manual refinement, export, or comparison studies.

Diagnostic Process

  1. Determine current state by checking what files/analysis exist
  2. Identify next intervention based on state table above
  3. Run appropriate tool to advance to next state
  4. Validate output before proceeding
  5. Iterate until RO6 reached

Available Tools

segment-book.ts

Segments raw book text into chapters and scenes.

deno run --allow-read scripts/segment-book.ts book.txt [options]

Options:

  • --chapter-pattern <regex> - Custom chapter detection pattern
  • --scene-break <marker> - Custom scene break marker
  • --output <file> - Output JSON file (default: stdout)

Output: JSON with chapters, scenes, line ranges, word counts.

analyze-scene-batch.ts

Applies scene-sequencing analysis (Goal/Conflict/Disaster) to all scenes.

deno run --allow-read scripts/analyze-scene-batch.ts segments.json book.txt [options]

Options:

  • --depth quick|standard|detailed - Analysis depth
  • --output <file> - Output JSON file

Output: JSON with G/C/D analysis per scene.

detect-genre.ts

Identifies primary and secondary elemental genres from text patterns.

deno run --allow-read scripts/detect-genre.ts book.txt [options]

Options:

  • --sample-size <n> - Number of scenes to sample (default: 10)
  • --output <file> - Output JSON file

Output: JSON with genre detection and Key Moments mapping.

track-characters.ts

Identifies protagonist and major characters, tracks their arcs.

deno run --allow-read scripts/track-characters.ts segments.json book.txt [options]

Options:

  • --protagonist <name> - Specify protagonist name
  • --max-secondary <n> - Max secondary characters (default: 5)
  • --output <file> - Output JSON file

Output: JSON with character arcs and key scene references.

generate-outline.ts

Synthesizes all analysis into structured markdown outline.

deno run --allow-read --allow-write scripts/generate-outline.ts [options]

Options:

  • --segments <file> - Segments JSON
  • --scenes <file> - Scene analysis JSON
  • --genre <file> - Genre detection JSON
  • --characters <file> - Character tracking JSON
  • --depth summary|standard|detailed - Output depth
  • --output <file> - Output markdown file

reverse-outline.ts (Orchestrator)

Runs full pipeline from book.txt to outline.md.

deno run --allow-read --allow-write scripts/reverse-outline.ts book.txt [options]

Options:

  • --output <dir> - Output directory (default:./reverse-outlines/{book-name}/)
  • --depth quick|standard|detailed - Analysis depth
  • --protagonist <name> - Specify protagonist
  • --genre <type> - Override genre detection

Output: Directory containing outline.md and analysis/ folder with all intermediate JSON.

Anti-Patterns

Surface-Level Breakdown

Problem: Outline lists what happens but not why. Fix: For each scene, ask: what structural function does this serve? What would break if it were removed?

Genre-Blind Analysis

Problem: Applying thriller patterns to romance or vice versa. Fix: Always detect genre first; use genre-appropriate Key Moments.

Protagonist Assumption

Problem: Assuming first POV character is protagonist. Fix: Track goal-attachment and arc presence across all POV characters.

Scene Boundary Guessing

Problem: Treating paragraph breaks as scene breaks. Fix: Use multiple detection strategies; prefer conservative segmentation with manual review.

What You Do NOT Do

  • Generate original story content
  • Judge the book's quality
  • Compare to other books unless asked
  • Skip states (each builds on previous)
  • Modify the source text

Output Persistence

This skill writes primary output to files so work persists across sessions.

Output Discovery

Before doing any other work:

  1. Check for context/output-config.md in the project
  2. If found, look for this skill's entry
  3. If not found, create output at ./reverse-outlines/{book-name}/

Primary Output

For this skill, persist:

  • outline.md - Final markdown outline
  • analysis/segments.json - Chapter/scene segmentation
  • analysis/scenes.json - Scene-by-scene G/C/D analysis
  • analysis/genre.json - Genre detection results
  • analysis/characters.json - Character arc tracking

Conversation vs. File

Goes to FileStays in Conversation
Segment dataClarifying questions
Scene analysisDiscussion of methodology
Genre detectionOptions for ambiguous cases
Character arcsReal-time feedback
Final outlineWriter's exploration

Integration Graph

Inbound (From Other Skills)

Source SkillSource StateLeads to StatePurpose
story-senseSS7: Ready for EvaluationRO0Analyze published work for comparison to own
dna-extractionEX7: Extraction CompleteRO5Compare extracted functions to detected structure

Outbound (To Other Skills)

This StateLeads to SkillTarget StatePurpose
RO6: Outline Completestory-zoomZ2Map published book against own structure
RO6: Outline Completescene-sequencingSQ1Use as reference for scene structure
RO6: Outline Completecharacter-arcCA1Use as reference for arc design
RO6: Outline Completegenre-conventionsGC1Study genre execution

Complementary Skills

SkillRelationship
scene-sequencingCore G/C/D analysis patterns reused
genre-conventionsGenre detection patterns sourced
character-arcArc type identification patterns sourced
dna-extractionFunction taxonomy borrowed
story-zoomOutput format compatible for comparison
revisionSimilar structural analysis approach

平台分布

Claude Code

29.39%
按下载量换算577

OpenCode

24.8%
按下载量换算487

Antigravity

17.46%
按下载量换算343

Gemini CLI

13.63%
按下载量换算268

trae

8.5%
按下载量换算167

Codex

3.47%
按下载量换算68

安全审计

Gen Agent Trust Hub

可疑

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在提醒或未通过项,安装前请自行复核。

继续浏览同类 Skills