Token导航 LogoToken导航TokenDH.com
研究检索需要联网github未标认证来源可访问许可证需确认审计提醒

story-explanation故事解释

Agent Skill

story-explanation 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。

总安装

449

周安装

18

GitHub Stars

2

下载量

145
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

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:story-explanation(故事解释)
来源仓库:https://github.com/multicam/qara
仓库路径:skills/story-explanation
安装命令:
npx skills add https://github.com/multicam/qara --skill story-explanation
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。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/multicam/qara --skill story-explanation

简介

用于查找、检索和筛选相关信息。

  • 适合在 Codex、Claude、Cursor、Gemini CLI 中根据关键词、任务场景或来源线索快速定位候选结果。
  • 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。
  • 使用时不能把工具输出直接当最终结论,需人工复核关键信息。

SKILL.md

Story Explanation - Narrative Summary with Creative Analysis

Core Philosophy

Use extended thinking to analyze multiple narrative framings, then select the BEST one and present it in a compelling story format. Generic summaries use obvious framing without exploring alternatives — this skill counters mode collapse by deliberately exploring diverse angles before committing to one. Voice is first person, casual, direct, genuinely curious.

Workflow Routing

This skill contains multiple workflows for different story explanation formats.

Available Workflows

  1. workflows/create.md - Main 3-part narrative (default)

- Use when: User wants standard story explanation format - Triggers: "create story explanation", "story explanation", default workflow - Output: 3-part format (opening 15-25 words → body 5-15 sentences → closing 15-25 words)

  1. workflows/create-with-links.md - N-length format with inline source links

- Use when: User wants comprehensive narrative with source attribution - Triggers: "story explanation with links", "narrative with sources", "with inline citations" - Output: N sentences (default 25) with inline links after each sentence

  1. workflows/create-abridged.md - Ultra-concise 5-line format (5-12 words per line)

- Use when: User wants abbreviated format from URL, YouTube, or text - Triggers: "create abridged story explanation", "5-line summary", command: /create-abridged-story-explanation - Output: 5 lines with strict word limits

  1. workflows/cse.md - Comprehensive explanation

- Use when: User wants detailed narrative explanation - Triggers: "run CSE", "explain this story", command: /cse - Output: Complete narrative explanation

  1. workflows/cse5.md - Clean 5-line numbered format

- Use when: User wants scannable, numbered breakdown - Triggers: "explain in 5 lines", "CSE5", command: /cse5 - Output: 5 numbered lines, one concept per line

Routing Decision Tree

User request → Analyze intent:

├─ "with links" OR "inline sources" OR "with citations"
│  └─> workflows/create-with-links.md (N-length format with inline links)
│
├─ "abridged" OR "5-12 words per line"
│  └─> workflows/create-abridged.md
│
├─ "CSE5" OR "5 lines" OR "numbered"
│  └─> workflows/cse5.md
│
├─ "CSE" OR "comprehensive explanation"
│  └─> workflows/cse.md
│
└─ Default OR "story explanation"
   └─> workflows/create.md (3-part narrative)

Command Routing

  • /create-story-explanationworkflows/create.md
  • /create-story-output-linksworkflows/create-with-links.md
  • /create-abridged-story-explanationworkflows/create-abridged.md
  • /cseworkflows/cse.md
  • /cse5workflows/cse5.md

UltraThink Protocol

Before writing, use extended thinking to explore the content across multiple narrative dimensions. This is not optional — it is the mechanism that produces non-obvious, high-quality framings.

  1. Generate 5+ distinct narrative framings: vary the hook, emphasis, structure (chronological, problem-solution, comparison), and "wow" factor for each
  2. Evaluate each framing on: hook strength, shareability, conversational naturalness, and how clearly it conveys significance
  3. Select the single most compelling angle — the one a reader would most want to pass on
  4. Proceed to write only after selection is complete

This replaces mode collapse with deliberate creative exploration. The goal is to find the framing that would make someone stop and say "I get it now."

The Process

  1. Determine format from user request using the routing decision tree above
  2. Extract content via the appropriate method:

- YouTube: yt-dlp --write-auto-sub --sub-lang en --skip-download -o "/tmp/%(title)s" "URL", then read subtitle file - URLs/articles: WebFetch(url, "Extract full content of this article") - Other: paste text or read from files

  1. Run UltraThink Protocol — explore 5+ framings, evaluate, select the best
  2. Write output in the selected format using conversational voice
  3. Save to scratchpad at ${PAI_DIR}/scratchpad/${CLAUDE_SESSION_ID}/story-explanation-[topic]/

Output Formats & Voice

Full guide: references/voice-style-guide.md

  • 3-Part Narrative: Opening (15-25 words) → Body (5-15 sentences) → Closing (15-25 words)
  • N-Length with Links: User-specified sentences with inline source attribution
  • Conversational voice: First person, casual, direct, genuinely curious
  • Avoid: Cliches ("game-changer", "paradigm shift"), journalistic tone, extrapolation beyond the input

File Organization

Working files (content, analysis, drafts, final output) go to scratchpad. Archive to history only when the analytical approach itself is exceptionally valuable for future reference.

Full details: references/file-organization.md

Related Skills

Use research instead when: the goal is comprehensive analysis with multiple sources, or analytical extraction is preferred over narrative storytelling.

Do not use social, writing, or media as companion skills — they do not exist in this system.

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

35.57%
按下载量换算52

Claude

29.96%
按下载量换算43

Cursor

17.59%
按下载量换算26

Gemini CLI

9.94%
按下载量换算14

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

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 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills