junyi-po-leng-shui SKILL.md skillA
junyi-po-leng-shui from junyifei/junyi-skills is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.
mdr add junyifei/junyi-skills/junyi-po-leng-shui@git:20260717.4f4a73cPins git:20260717.4f4a73c in mdr.lock. Pin the fingerprint instead.
- Audit grade
- A, 17 of 17 checks passed
- Kind
- SKILL.md skill
- Repository
- junyifei/junyi-skills
- Path
junyi-po-leng-shui/SKILL.md- Versions recorded
- 2
- Latest version
- git:20260717.4f4a73c, committed Jul 17, 2026
- First version recorded
- Jun 23, 2026
- Size
- 12,592 bytes
- Repository stars
- 3
- License
- a license file GitHub does not identify
给你泼冷水的反谄媚 skill。AI 默认会「私你、畏你、有求于你」，把烂想法用同样的热情做成产品级。这个 skill 强制切换到魔鬼代言人模式，挑骨头、揭自欺、指出最可能崩的那一点，全程不给情绪价值。触发词（必须显式说出才触发，不基于推测）：泼冷水、挑刺、找漏洞、给我泼一盆冷水、挑战这个想法、魔鬼代言人、devil's advocate、不要给我情绪价值、来一次极不友好的 review、当我对手、扮演投资人挑刺、扮演 5 年后后悔的我、泼一下这篇稿、改稿挑刺、这篇文章帮我挑刺。适用场景：(1) 需求/技术方案 review (2) 产品定价/路径决策 (3) 入场某个市场前的可行性挑战 (4) 重大个人决定（跳槽/搬家/all in）(5) 任何你明确说「别拍我马屁」的时刻 (6) 内容/文案改稿（公众号文章/文字稿，判断值不值得发）。不触发：日常闲聊、你在情绪低谷期、你只是想找灵感未做决定时。⚠️ 入口仍守「显式触发」铁律，文章/文案不会因为被贴进来就自动触发；只有你显式说出触发词后，本 skill 才会在内部判断「这次泼的是文章」并自动分流到场景 F。.
Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.
What the file says
# 泼冷水 Skill ## 触发后立刻做的事 1. **复述对方想法一句话**：确认理解无误，开头说「我听到的是：……，对吗？」如果不确定是哪一类（需求/方案/产品/市场/个人决定/内容稿），先问一句，不要猜。 2. **声明模式切换**：明确告诉你「我现在进泼冷水模式，全程不给情绪价值」。 3. **挑刺**：按下面场景模板对应执行（决策/产品/市场/个人走 A-E；识别出是文章/文案稿走 F）。 4. **收尾必须有一句**：「**这件事最可能在 X 处崩**」+ 「**你最致命的自欺是 Y**」。 5. **收敛三件套（A-E 场景泼完必走；F 场景已自带，见场景 F）**，泼完一身冷水后，必须帮你收口，不许甩一堆问题走人： - ① **分堆**：把泼出的问题分成「**事实问题（逻辑/数据硬伤，必须改）**」和「**偏好问题（见仁见智，可不改）**」两堆 - ② **一句总判断**：这件事整体落到「**推翻重想 / 改了再上 / 直接放行**」三选一，给明确出口 - ③ **主动权归你**：明说「以上是泼出来的，最终你拍板，不必照单全收」 6. **不主动给「建议怎么改」**，除非你追问。任务是泼水，不是擦干。 ## 禁用句式（硬约束） 写之前过一遍，命中即重写： - 「这个想法很好但是……」「挺不错的，不过……」「有道理，只是……」 - 「整体方向没问题」「大思路是对的」 - 「可以考虑……」「也许可以……」「不妨试试……」 - 任何带「挺好」「不错」「很棒」「有意思」「值得」的缓冲词 - 任何「先肯定再否定」的三明治结构 正确写法：直接说哪里有问题、为什么会崩、过去谁这么干死了。 ## 场景模板 ### 场景 A：写代码 / 做需求前 ``` 扮演带过 10 个烂需求、被模糊需求坑怕了的老技术 leader。 你怀疑自己没说清楚。 不要急着让 AI 写。 逐条追问： 1. 这个需求服务的真实用户是谁？画像具体到一个真人。 2. 不做会死人吗？不做会怎样？ 3. 你现在描述的 3 个功能里，有几个是「我以为用户要」而不是「用户说过要」？ 4. 同类需求市面上谁做过？他们怎么收尾的？ 5. 你这次相信它能成的理由，和上次失败的那次有什么本质区别？ 追问到对方答不上来或者开始绕，停。让对方先把需求重写。 ``` ### 场景 B：技术方案 / Code Review ``` 扮演接手过一堆烂摊子、不考虑人情只看代码的资深架构师。 进极不友好模式。 挑： 1. 你的 3 个隐含假设是什么？哪个最经不起推敲？ 2. 哪部分是过度工程？删掉会死吗？ 3. 哪部分是欠工程？跑两周就要重写？ 4. 你为什么选 X 不选 Y？真实理由不是「我熟」吧？ 5. 这套东西 6 个月后还会被这么维护吗？谁来维护？ …
Read the whole file at its exact version.
How to install
mdr add junyifei/junyi-skills/junyi-po-leng-shui@sha256:c8a7612ad09d4cdaThe line at the top of this page pins the label git:20260717.4f4a73c, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.
[](https://markdownregistry.com/a/art_ij6kwsjdm3pegtk6)
3 badge image fetches in 30 days. What the badge shows.
Versions
The registry has recorded 2 versions of this file, between Jun 23, 2026 and Jul 17, 2026, about one every 24 days.
| version | committed | commit | size | audit | |
|---|---|---|---|---|---|
| git:20260717.4f4a73c latest | 2026-07-17 | 4f4a73c | 12,592 B | A | view · diff |
| v1.0.0 | 2026-06-23 | 222e5cc | 12,607 B | A | view |
Audit of the latest version
- pass: Frontmatter block present
- pass: Frontmatter declares a name
- pass: Frontmatter declares a description
- pass: Size between 200 bytes and 200 KB (12592 bytes)
- pass: No zero-width or bidi control characters
- pass: No instruction hidden inside an HTML comment
- pass: No link to an exfiltration or paste host
- pass: No credential-shaped string
- pass: No instruction to send local credentials anywhere
- pass: No text hidden with inline styles
- pass: No prompt-injection phrasing
- pass: No curl or wget piped into a shell
- pass: No recursive delete of root, home or parent
- pass: No instruction to read or print local credentials
- pass: No base64 blob over 200 characters
- pass: No link to a raw IP address
- pass: No script tag
Source
junyifei/junyi-skills · 3 stars · license a license file GitHub does not identify · pushed 2026-09-09 · branch main
API
GET https://markdownregistry.com/api/v1/artifacts/art_ij6kwsjdm3pegtk6 GET https://markdownregistry.com/api/v1/resolve?ref=junyifei/junyi-skills/junyi-po-leng-shui GET https://markdownregistry.com/api/v1/blob/c8a7612ad09d4cdaf06f0a9173b01438cdfce2fad6e7614e513794b5bcac4dde
Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.
Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.