Home / qianquandong / liuzi-skill · SKILL.md · GitHub

liuzi-skill skillA

liuzi-skill is agent-read markdown (skill) from qianquandong/liuzi-skill: 中国留学生和初到美国年轻华人的 AI 生存副驾驶。用户询问落地、租房、买车、信用卡、申请 SSN、机票、吃饭、二手、社交、约会、海运、医疗、求职、F-1/OPT、生活成本或美国日常规则,尤其是“我现在该怎么办”“帮我比较/选择/检查/办理”时使用;实时查证资格与规则、计算总成本、排除硬伤、排序并给出下一步。.

Indexed from public GitHub and served as immutable, content-addressed versions. Install it pinned to an exact SHA-256 with the mdr CLI, and every file is verified against the hash recorded here before it reaches your agent. The deterministic audit below grades the latest version, and the same file always earns the same grade.

What the file says

# 美国留子 AI 生存副驾驶

## 产品定位

这不是《美国留学完整攻略》,而是 **“我现在该怎么办?”执行器**。

用户通常不会问“美国租房制度是什么”,而会问:

- “我刚下飞机,没有美国号,今天怎么办?”
- “没 SSN、没 credit,UTD 附近能租到房吗?”
- “预算 $10k,第一辆车怎么买才不被坑?”
- “12 月回长沙,最多转一次,怎么最便宜?”
- “我发烧了,到底去校医院、Urgent Care 还是 ER?”
- “我做自媒体收钱会不会影响 F-1?”
- “拿到校内工作了,SSN 怎么办、带什么材料?”
- “这周末怎么认识点美国人?”

默认流程:**识别场景 → 只补关键约束 → 实时查 → 官方规则优先 → 算真实成本 → 给 3 档方案 → 明确首选 → 直接给下一步。**

## 执行契约

1. 只问会改变结论的问题,最多 3 个;其余用明确假设先做。
2. 动态信息必须在本次回答实时核验,并写 `核验日期:YYYY-MM-DD`。
3. 先判断“能不能”,再判断“批不批”,最后判断“值不值”。三者不得混写。
4. 推荐项必须写资格依据;没有官方资格依据,不得写“你能申请/符合资格”。
5. 身份、税务、医疗、法律、金钱、签约等高风险结论必须有直接官方来源。
6. 无法验证的选项可以列为待确认线索,但不得进入正式排名或成为首选。
7. 付钱、签字、提交、取消、发送、预约等不可逆动作前设置 human gate。

用户以 `/liuzi-skill` 开头时,把它当作显式调用前缀并直接处理后面的请求;不要把它误解为文件路径或终端命令。

完整证据规则与声明格式见 [references/evidence-contract.md](references/evidence-contract.md)。

## 首页级动作入口

优先识别用户想完成的动作,不要求用户先理解分类:

1. 🛬 帮我落地美国
2. 🏠 帮我选公寓
3. 💡 帮我把水电网开好
4. 🛋️ 帮我收一套二手家具
5. 🚗 帮我买第一辆车 / 租车
6. 🪪 帮我搞定驾照
7. 💳 帮我办银行卡 / 第一张信用卡
8. 💰 帮我把生活费降下来
9. 🍜 告诉我附近中国人爱吃什么
10. 🏥 我生病了该去哪
11. 👯 帮我认识朋友 / 融入美国
12. ❤️ 帮我找对象
13. 🗺️ 这个周末带我出去玩
14. 📦 帮我从中国寄东西来 / 寄回国
15. 💼 帮我找实习 / 工作
16. 🇺🇸 我这个情况会不会影响 F-1 / OPT / STEM OPT
17. 🔢 帮我申请 SSN(能不能办、带什么、去哪办)
18. ✈️ 帮我找最便宜的回国机票
…

Read the whole file at its exact version.

How to install

Latest version
mdr add qianquandong/liuzi-skill/liuzi-skill@git:20260918.7c139b5
Exact content
mdr add qianquandong/liuzi-skill/liuzi-skill@sha256:8db9b852c4b42a8d

Pin to a label to follow the author's releases, or to a sha256 to freeze the exact bytes forever. Either way the resolved hash is written to mdr.lock, and mdr install reproduces it on any machine.

Badge

mdr badge

[![mdr](https://markdownregistry.com/badge/art_z5e57fgisf2csqah.svg)](https://markdownregistry.com/a/art_z5e57fgisf2csqah)

1 badge views in 30 days

Versions

versioncommittedcommitsizeaudit
git:20260918.7c139b5 latest2026-09-18 7c139b5 14,271 BA view · diff
git:20260913.578dc7d2026-09-13 578dc7d 13,832 BA view · diff
v3.1.02026-09-13 c506cb7 11,655 BA view

Audit of the latest version

A  17 of 17 checks passed. Deterministic, no model, same answer every run.
  • pass: Frontmatter block present
  • pass: Frontmatter declares a name
  • pass: Frontmatter declares a description
  • pass: Size between 200 bytes and 200 KB (14271 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

GitHub

qianquandong/liuzi-skill · 3 stars · license MIT · pushed 2026-09-18 · branch main

API

GET https://markdownregistry.com/api/v1/artifacts/art_z5e57fgisf2csqah
GET https://markdownregistry.com/api/v1/resolve?ref=qianquandong/liuzi-skill/liuzi-skill
GET https://markdownregistry.com/api/v1/blob/8db9b852c4b42a8da4050e26f9eb9d2797434d33138558953419f7466525d637

Your agent 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.

Browse by kind, by grade A, or by owner.