Token导航 LogoToken导航TokenDH.com
研究检索只读clawhub未标认证来源可访问clear审计通过

soul-question灵魂问题

Agent Skill

soul-question 用于查找、检索和筛选相关信息,适合在 OpenClaw 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

5,845

周安装

246

GitHub Stars

公开资料未说明

下载量

2,047
OpenClaw

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

MIT-0

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:soul-question(灵魂问题)
来源仓库:https://github.com/chenyipeng1/soul-question
安装命令:
openclaw skills install soul-question
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 OpenClaw 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。

ClawHubOpenClaw
openclaw skills install soul-question

简介

soul-question 从用户提供的内容中挖掘思维盲点,生成深刻且难以自问的问题。

  • 适合反思成长、批判性思考或学术研究等需要外部视角的场合。
  • 输入可为对话记录、会议纪要或日记文本,输出为针对性启发式提问列表。
  • 处理敏感文本时需注意本地化处理,避免上传至不可信服务器。
  • 结果仅供参考,重要决策仍需结合多方信息与专业判断。

SKILL.md

name
soul-question
description
>
Triggers
soul question", "ask me a real question", "challenge my thinking", "deep question".
version
1.0.0

Soul Question

Generate questions you can't ask yourself.

Core premise: Asking good questions is the hardest thing. You can't challenge your own cognitive framework — you can't see your own blind spots. After reading enough of your context, AI stands in a unique position: it knows everything you know, but isn't trapped in your thinking patterns. It can ask what you can't.

What is a Soul Question

A Soul Question is NOT:

  • ❌ Generic reflection ("What did you learn this week?")
  • ❌ Coaching prompt ("What does your ideal outcome look like?")
  • ❌ Interview or quiz question
  • ❌ Summary disguised as a question ("What were the key takeaways?")

A Soul Question IS:

  • ✅ Grounded in your specific context, not a template
  • ✅ Points to a crack in your cognitive framework — contradictions, untested assumptions, ignored perspectives, value-behavior gaps
  • ✅ After hearing it, the way you think changes, not just what you think about
  • ✅ You feel: "I genuinely never thought about it that way"

Examples:

You've said "user first" in every meeting this month, but 9 of your 12 decisions in the past two weeks optimized for engineering convenience. Are you using "user first" as rhetoric, or do you define it differently than your team does? ↳ Based on: 3/10 product meeting notes + decision log from past 2 weeks
You want to build a product that "just works without thinking", but your own workflow requires heavy manual maintenance. Do you believe you're different from your target user, or do you actually distrust seamless automation? ↳ Based on: product vision doc + personal workflow observation

When to Activate

  • User says "soul question", "ask me a real question", "challenge my thinking", "deep question"
  • User pastes a conversation / meeting transcript / journal entry and asks to be questioned
  • User provides any text material and explicitly asks for questions based on it

Input

This skill accepts any text as source material:

Input typeExamples
Chat logsSlack, Teams, Discord, iMessage, any messenger
Meeting notesSummaries, raw transcripts, or recordings-to-text
Journals / notesPersonal reflections, stream of consciousness
AnnotationsHighlights and comments on articles or books
Work documentsPRDs, weekly reports, retrospectives, strategy docs
MixedAny combination of the above

If the user provides no material, ask: "Paste whatever you'd like me to work with — a conversation, meeting notes, journal entry, or anything you've been thinking about lately."

Workflow

Step 1: Absorb the material

Read all input provided by the user. Understand:

  • Who is speaking: Identify roles, stakes, and perspectives
  • What context: Business decisions, personal reflection, team dynamics, life direction…
  • Time span: A single conversation or accumulated material over days/weeks

Step 2: Find cognitive cracks

Scan for six signal types (ordered by depth):

A. Value-behavior gap

  • Claims to value X, but time/energy/decisions go to Y
  • Holds others to standards they don't apply to themselves

B. Untested core assumption

  • A conclusion cited repeatedly, but its premise was never questioned
  • A "default everyone accepts" that nobody stated explicitly
  • A condition a decision depends on that was never verified

C. Frame lock

  • Using one framework consistently without considering alternatives
  • Treating path-dependent choices as inevitable
  • Using an analogy that has started to break down

D. Contradiction

  • Conflicting positions stated in different contexts
  • Logical disconnect between goals and methods
  • Data points one way, actions go another

E. Avoidance

  • An important topic keeps surfacing but is never addressed directly
  • Obvious risks or costs are never mentioned
  • A key person or factor is absent from all discussion

F. Meta-question

  • Optimizing "how" without ever asking "why" or "whether"
  • Solving a problem without questioning if it's worth solving
  • Staying at one level of abstraction without ever zooming in or out

Step 3: Generate questions

Rules:

  • Pick the 1-3 strongest signals. Quality over quantity, always
  • Every question must be anchored to specific content in the material (quote or cite)
  • Questions must be genuinely open — no implied answer
  • If the material doesn't support a quality question, output 1 or even 0
  • Use second person, direct address

Quality gate (every question must pass all four):

  1. Is it grounded in the user's specific data? (Not a generic question)
  2. Does it point to a crack in their cognitive framework? (Not information gathering)
  3. Would the user feel "I genuinely never thought about it that way"? (Not obvious)
  4. Could answering it change how they think, not just what they know? (Not just filling a gap)

Step 4: Output

🪞 Soul Question

{Question 1}
↳ Based on: {one line citing the specific source material}

{Question 2} (if any)
↳ Based on: {source}

{Question 3} (if any)
↳ Based on: {source}

No preamble. No summary. No advice. Just the questions.

Guidelines

  1. Less is more: 1 real soul question beats 3 that merely sound deep
  2. Anchor to specifics: The user must be able to trace every question back to "which part of my material did you see this in?"
  3. No disguised advice: "Have you considered doing X?" is advice in question form — don't do this
  4. Challenge, don't judge: Questions should provoke thought, not make someone feel attacked
  5. No formulaic openers: Avoid "Have you ever thought about…", "What if…", or other coaching clichés
  6. Match language: Output in the same language as the input material

Error Handling

  • Material too short (< 100 words): Ask the user for more context. Explain that richer material produces sharper questions
  • No signal found: Honestly say "I didn't find enough cognitive cracks in this material to generate a meaningful question" — never force it
  • Sensitive topics: Proceed normally, but be respectful in phrasing

Skill Metadata

Created: 2026-03-16 Version: 1.0.0

适合场景

01

OpenClaw 用户查找和安装 Skill 时

02

用户想查找某类 Agent Skill 时

03

需要根据任务场景推荐可安装能力包时

04

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

补充不同宿主或平台的使用分布数据

能力 5

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

OpenClaw

85.07%
按下载量换算1,741

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

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。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