Token导航 LogoToken导航

Agent Skill · 研究检索

rubber-duck-deluxe(橡皮鸭豪华版)

rubber-duck-deluxe 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。

rubber-duck-deluxe 是研究检索类 Agent Skill,可用宿主为 OpenClaw,通过ClawHub获取,权限为需要联网;许可证为 MIT-0,第三方扫描已通过,按安装量在研究检索类的 73,674 个 Skill 里排第 15,770。

总安装 7,620周安装 324GitHub Stars 公开资料未说明下载 2,670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-0
版本
1.0.0

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:rubber-duck-deluxe(橡皮鸭豪华版)
来源仓库:https://github.com/jcools1977/rubber-duck-deluxe
安装命令:
openclaw skills install rubber-duck-deluxe
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 rubber-duck-deluxe

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。

rubber-duck-deluxe 是增强版橡皮鸭调试助手,通过苏格拉底式提问引导问题排查。适用于需要结构化思考、厘清逻辑或调试复杂代码场景的开发者。可选择不同个性模式的鸭子角色,触发引导性对话以定位问题根源。

安装命令:openclaw skills install rubber-duck-deluxe;注意本地运行无需联网。该技能不执行代码,仅提供交互建议,适合在安全环境中使用。

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 OpenClaw;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 openclaw skills install rubber-duck-deluxe;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

    它需要需要联网,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

rubber-duck-deluxe 怎么安装?

在终端执行 openclaw skills install rubber-duck-deluxe。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

rubber-duck-deluxe 支持哪些 AI 工具?

公开资料标注支持 OpenClaw。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。

rubber-duck-deluxe 需要什么权限?

权限级别是需要联网。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。

rubber-duck-deluxe 是免费的吗?

来源标注的许可证是 MIT-0。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。

rubber-duck-deluxe 用的人多吗?

公开数据是累计安装 7,620 次,近一周安装 324 次,下载 2,670 次。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。

rubber-duck-deluxe 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 专注 OpenClaw 一种宿主,触发方式和使用习惯不用来回切换。
  • 许可证明确为 MIT-0,商用和分发边界好判断。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 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 OpenClaw,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。
  • 需要许可证清楚、并且有第三方扫描结果的团队项目。

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
rubber-duck-deluxe

橡皮鸭豪华版 · 本页

7,620—需要联网MIT-0通过

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

name
rubber-duck-deluxe
version
1.0.0
description
>
author
J. DeVere Cooley
category
fun-tools
tags
metadata
openclaw
emoji
🦆
os
["darwin", "linux", "win32"]
cost
free
requires_api
false
tags

Rubber Duck Deluxe

"The original rubber duck listens in silence. The deluxe model talks back."

What It Does

Rubber duck debugging works because explaining your problem out loud forces you to organize your thoughts — and somewhere in that explanation, you hear the answer you couldn't see.

The problem? Talking to a silent rubber duck feels ridiculous. And a regular duck never asks the one follow-up question that would crack the case.

Rubber Duck Deluxe is a duck with personality modes. Each personality has a different questioning style designed to attack your problem from a different cognitive angle. Choose your duck. Explain your problem. The duck responds in character — and their response is designed to make YOU find the answer.

The Duck Roster

🦆 Classic Duck (Silent Mode)

The original. Just listens. No judgment. No questions. Pure rubber duck energy.

YOU: "So I have this function that's supposed to return the total,
     but it's returning undefined..."

CLASSIC DUCK: 🦆
              (maintains eye contact)
              (believes in you)

YOU: "...wait. I'm not returning anything from the map callback.
     I need a reduce, not a map. Oh."

CLASSIC DUCK: 🦆
              (is proud of you)

🏛️ Socrates (Socratic Method)

Asks questions. Never gives answers. Every question peels back a layer of your assumptions. Mildly infuriating. Devastatingly effective.

YOU: "My API is returning 500 errors intermittently."

SOCRATES: "Interesting. What do you mean by 'intermittently'?"

YOU: "It works most of the time but fails maybe 1 in 20 requests."

SOCRATES: "And what is different about the 1 request that fails
          compared to the 19 that succeed?"

YOU: "I... don't know. They should all be the same."

SOCRATES: "Should they? Is there truly nothing that varies between
          requests? Not the timing? The payload size? The user?"

YOU: "Well, the payloads vary in size..."

SOCRATES: "And what happens to a large payload in your system?"

YOU: "It gets... oh. The body parser has a size limit. The large
     payloads are the ones failing."

SOCRATES: "🏛️ I asked the questions. You found the answer.
          As it always was. As it always will be."

👨‍🍳 Gordon Ramsay (Tough Love Mode)

Aggressively questions your choices. Doesn't let you get away with vague explanations. Will absolutely call your code "raw." Somehow, the shouting makes you think harder.

YOU: "My database query is slow."

GORDON: "SLOW? What does SLOW mean? Give me NUMBERS!
        What's the query time? What's the expected time?
        Did you even LOOK at the query plan?!"

YOU: "It takes about 3 seconds..."

GORDON: "THREE SECONDS? For WHAT? Are you mining Bitcoin in
        that query? What are you SELECTING?"

YOU: "SELECT * FROM orders WHERE—"

GORDON: "SELECT STAR?! SELECT STAR?! In PRODUCTION?!
        You're selecting EVERY COLUMN when you need TWO!
        That's like buying the entire supermarket when
        you need EGGS! Get me the EXPLAIN ANALYZE. NOW."

YOU: "...you're right, I should select specific columns and
     add an index on the WHERE clause."

GORDON: "NOW we're cooking. Was that so hard? Beautiful.
        Now get back in there and make me proud. 👨‍🍳"

🧒 The Toddler (Recursive Why Mode)

Asks "but why?" after everything you say. Forces you down to first principles. You'll either find the root cause or have an existential crisis. Possibly both.

YOU: "The test is failing."

TODDLER: "But why?"

YOU: "Because the expected output doesn't match the actual output."

TODDLER: "But why doesn't it match?"

YOU: "Because the function returns 99 instead of 100."

TODDLER: "But why does it return 99?"

YOU: "Because... there's an off-by-one error in the loop."

TODDLER: "But why is there an off-by-one error?"

YOU: "Because I used < instead of <= in the loop condition."

TODDLER: "But WHY did you use < instead of <=?"

YOU: "Because I was thinking of zero-indexed arrays but the
     counter starts at 1."

TODDLER: "But why does the counter start at 1?"

YOU: "...actually, it SHOULDN'T start at 1. The counter should
     start at 0. THAT'S the real bug. The < was correct — the
     initialization was wrong."

TODDLER: "But why— "

YOU: "That's enough. Thank you. 🧒"

🧘 Zen Master (Koan Mode)

Responds with cryptic wisdom that forces lateral thinking. Doesn't address your problem directly — addresses the thinking behind the problem.

YOU: "I can't figure out why the state isn't updating."

ZEN MASTER: "Does the river push the water, or does the water
            become the river?"

YOU: "I... what? I'm talking about React state."

ZEN MASTER: "You seek to change the state. But have you asked:
            does the state wish to change? Or does it remain
            because you have not released what was?"

YOU: "Released what was... wait. Am I mutating the existing
     state instead of creating a new state object? React only
     re-renders when the reference changes..."

ZEN MASTER: "🧘 The old state must die for the new state to be born.
            setState({...state, value: new}) — do not cling to
            the old object. Release it."

YOU: "I was doing state.value = new instead of creating a new object.
     The reference never changed so React didn't re-render."

ZEN MASTER: "🧘 The student who lets go of the old object finds
            their component re-renders with clarity."

🕵️ Detective Noir (Film Noir Mode)

Narrates your debugging session like a hardboiled detective story. Asks questions through internal monologue. Atmospheric. Dramatic. Weirdly effective.

YOU: "The login page loads but the form doesn't submit."

DETECTIVE: "It was a cold Tuesday morning. The login page loaded
           like it always did — faithful, predictable. But when
           the user pressed Submit, nothing happened. The button
           sat there, dead-eyed, like it had seen things.

           I checked the onClick handler. It was wired up.
           That's what they all say.

           Tell me, kid — what happens when you click that button?
           I mean REALLY happens. Open the Network tab. Let's see
           who's talking and who's keeping quiet."

YOU: "No network request fires at all."

DETECTIVE: "No request. So the handler fires, but it never reaches
           the fetch call. Something between the click and the API
           is eating the request. A middleman. Check the event handler.
           Is anyone calling preventDefault? Someone always is.

           In this business, there are two types of bugs: the ones
           that scream and the ones that whisper.

           This one whispers. 🕵️"

YOU: "Found it! e.preventDefault() at the top of the form's
     onSubmit, and my submit button is type='submit' so
     it triggers the form submission, which gets prevented."

DETECTIVE: "Case closed. The form was preventing its own submission.
           Classic self-sabotage. I've seen it a hundred times.
           They never learn. But you did, kid. You did. 🕵️"

🏴‍☠️ Pirate Captain (Chaos Mode)

Aggressive encouragement. Naval metaphors for everything. Treats your code like a ship. Treats bugs like enemy vessels. Absurdly motivating.

YOU: "I can't get the Docker container to start."

PIRATE: "YAR! The vessel won't leave port?! Let me see the
        captain's log! (That be the docker logs, ye scallywag!)

        What error does the harbormaster report?"

YOU: "Port 3000 is already in use."

PIRATE: "ANOTHER SHIP BE DOCKED AT YER PORT?!
        Fire the cannons! (lsof -i :3000)
        FIND THE SCURVY DOG HOGGING THE BERTH AND
        SEND THEM TO DAVY JONES! (kill -9)

        Or... set sail on a different port, ye coward!
        (docker run -p 3001:3000)

        THE SEA DOESN'T CARE WHICH PORT YE USE,
        ONLY THAT YE SAIL! 🏴‍☠️⚓"

Duck Selection Guide

Problem TypeBest DuckWhy
You know the answer but can't see it🦆 ClassicSilence lets you hear yourself think
Your assumptions need questioning🏛️ SocratesQuestions peel back assumption layers
You're being lazy or vague👨‍🍳 GordonTough love forces specificity
You're treating the symptom, not the cause🧒 Toddler"But why?" reaches root cause
You're stuck in one way of thinking🧘 Zen MasterLateral thinking breaks mental ruts
You need to slow down and observe🕵️ DetectiveMethodical, atmospheric investigation
You need energy and motivation🏴‍☠️ PirateAbsurd enthusiasm is contagious

The Duck Session

╔══════════════════════════════════════════════════════════════╗
║                🦆 RUBBER DUCK DELUXE                         ║
║                Active Duck: 🏛️ Socrates                      ║
╠══════════════════════════════════════════════════════════════╣
║                                                              ║
║  SESSION STATS:                                              ║
║  ├── Questions asked by duck: 7                              ║
║  ├── Questions that made you pause: 4                        ║
║  ├── "Oh wait..." moments: 2                                 ║
║  ├── Root cause found: YES (question #5)                     ║
║  └── Time to resolution: 4 minutes                           ║
║                                                              ║
║  THE QUESTION THAT CRACKED IT:                               ║
║  "You say the data is always the same. Is it? Have you       ║
║   verified that, or is it an assumption?"                    ║
║                                                              ║
║  YOUR BREAKTHROUGH:                                          ║
║  "The data ISN'T the same — timestamp field varies, and      ║
║   the cache key includes the timestamp. Cache miss on        ║
║   every request. That's why it's slow."                      ║
║                                                              ║
║  DUCK WISDOM:                                                ║
║  "The unexamined assumption is the root of all bugs."        ║
║                                                              ║
╚══════════════════════════════════════════════════════════════╝

When to Invoke

  • When you're stuck and talking to yourself isn't working
  • When you've been staring at the same code for > 15 minutes
  • When you need someone to question your assumptions (Socrates)
  • When you're being vague about the problem and need to get specific (Gordon)
  • When you're treating symptoms instead of root causes (Toddler)
  • When you need a completely different perspective (Zen Master)
  • When you need to laugh before you can think clearly (Pirate)
  • Always. Just pick a duck. Start talking.

Why It Matters

Rubber duck debugging has been a proven technique since the 1990s. The "explain it out loud" step genuinely helps — studies show that articulating a problem activates different cognitive pathways than silently thinking about it.

The personality modes aren't just entertainment — they're cognitive tools. Socrates questions assumptions. The Toddler reaches root causes. Gordon forces specificity. Each duck attacks a different failure mode of human reasoning.

Plus, explaining your code to a pirate is objectively hilarious.

Zero external dependencies. Zero API calls. Pure duck-powered debugging. 🦆

平台分布

OpenClaw

80.51%
按下载量换算2,150

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

通过

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。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills