← Back to Skills Marketplace
cp3d1455926-svg

friend-skill

by cp3d1455926-svg · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ Security Clean
67
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install friend-skill
Description
创建和管理基于你们私密聊天数据的朋友数字分身,实现朋友不在线时的温暖陪伴和模拟对话体验。
README (SKILL.md)

🤝 朋友 Skill · 数字分身制造术

让你的朋友住进你的AI里。

🌟 这是什么?

把女娲Skill的「人格提炼」和Yourself-Skill的「私人数据注入」合并,专为朋友场景优化的Skill。

它不代替你的朋友,而是在朋友不在线时,给你一个温暖的陪伴。

🚀 快速开始

1. 创建你的第一个朋友分身

cd friend-skill
node scripts/friend-builder.js

按提示输入朋友的信息,等待几秒,你就拥有了一个数字朋友!

2. 和朋友聊天

node scripts/chat-demo.js 张三
# 把"张三"换成你朋友的名字

🎮 使用场景

场景 适合
😂 吐槽日常 想找朋友吐槽但TA不在线
💡 征求意见 模拟朋友对某事的看法
📝 练习对话 不知道怎么说出口的话,先和数字朋友排练
🎭 角色扮演 用朋友的视角看问题
🗃️ 回忆录 记录和朋友的共同记忆,不会忘记

📂 项目结构

friend-skill/
├── SKILL.md                      # Skill定义文件
├── scripts/
│   ├── friend-builder.js          # 朋友分身构建器(核心)
│   └── chat-demo.js              # 聊天测试器
├── friends/                      # 生成的朋友分身文件(自动创建)
│   └── 张三-friend.json          # 示例输出
└── references/
    └── friend-template.json      # 朋友数据模板

🧩 进阶用法

从聊天记录文件导入

把你和朋友的聊天导出为 .txt 文件(每行一条消息),构建器会自动分析:

# 在交互模式中选择导入聊天记录
# 或者直接用JSON构建
node scripts/friend-builder.js --build data.json

JSON构建格式

{
  "name": "张三",
  "records": [
    "今天好烦啊",
    "哈哈哈笑死我了",
    "一起去吃饭吗"
  ],
  "personality": {
    "traits": ["搞笑", "暖男"]
  },
  "relationship": {
    "role": "损友",
    "yourName": "小老弟"
  },
  "interests": {
    "love": ["游戏", "电影"],
    "hate": ["数学", "早起"]
  },
  "voice": {
    "greeting": "你咋来了",
    "laugh": "哈哈哈你认真的?"
  }
}

🔒 隐私保护

这是最重要的部分!

  • 所有数据在本地处理 —— 朋友的信息不会上传到任何服务器
  • 聊天记录只用于分析 —— 分析完后可以删除原文件
  • 你控制一切 —— 生成的分身文件你可以随时删除
  • ⚠️ 只导你和朋友的对话 —— 不要导入群聊或包含其他人隐私的记录

🧠 设计哲学

朋友Skill的核心不是「完美模拟」——因为真正的朋友是不可替代的。

它只做三件事:

  1. 记住你们之间的温暖 —— 共同回忆、专属梗、互相的称呼
  2. 模仿朋友说话的方式 —— 语气、口癖、反应习惯
  3. 在你需要的时候出现 —— 不是替代,而是补充

📖 和女娲Skill的区别

维度 女娲Skill 朋友Skill
数据源 公开资料+网络搜索 你和朋友的私密数据
目标 思维框架模拟 关系温度模拟
公开性 可公开分享 仅你自己可用
需要 大量公开资料 少量聊天记录+你的回忆
输出 思维顾问 聊天伙伴

📝 TODO

  • 情感分析增强(识别用户情绪并匹配朋友的反应模式)
  • 长期记忆(记住上次聊到哪了)
  • 多模态支持(朋友的语音/表情包风格)
  • 自动「刷新」(根据新聊天记录更新人格)
  • 图形界面(可选)

友情提示:别让你的真朋友发现你有个AI版TA...除非TA也觉得这很酷😎

Usage Guidance
This package appears to operate locally and match its description, but you should still: 1) inspect the full contents of all shipped JS files (the provided dump showed a truncation) to confirm there are no network calls or hidden modules; 2) run the scripts in an isolated environment (e.g., a disposable VM or container) the first time; 3) avoid importing group chats or third-party private data as the README warns; 4) verify Node version and run with least privilege (no elevated rights); and 5) delete generated persona files when you no longer want them or keep them encrypted if they contain sensitive information.
Capability Assessment
Purpose & Capability
Name/description (create/manage friend digital doubles from private chat data) match the code and SKILL.md. The code manipulates local chat records, extracts tone/frequency features, generates persona JSON, and provides local chat tooling — all coherent with the stated purpose.
Instruction Scope
SKILL.md tells the user to run local Node scripts (friend-builder.js, chat-demo.js) and to import local .txt/.json chat exports. The scripts shown read/write files under the project (friends/), use only filesystem and readline, and do not instruct reading unrelated system files or contacting remote endpoints.
Install Mechanism
No install specification (instruction-only plus shipped JS files). That is low-risk and consistent with a Node-based local utility; nothing is downloaded or installed automatically.
Credentials
The skill declares no required environment variables, credentials, or config paths. The code uses only fs/path/readline and stores persona JSON locally — requested permissions are proportional to functionality.
Persistence & Privilege
Flags: always=false and the skill does not request persistent platform privileges. It reads/writes its own files under the project (friends/), which is expected for this purpose and does not modify other skills or system-wide agent settings.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install friend-skill
  3. After installation, invoke the skill by name or use /friend-skill
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
friend-skill 1.0.0 - 首次发布,结合女娲Skill的人格提炼和Yourself-Skill的私人数据注入,专为朋友陪伴场景设计 - 支持通过脚本快速创建和体验数字朋友分身 - 聊天记录和JSON文件多种数据导入方式 - 注重隐私保护,所有数据本地存储与处理 - 个性化模拟朋友的语言风格、兴趣与关系特征 - 已包含多种常见朋友使用场景和后续待开发功能预告
Metadata
Slug friend-skill
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is friend-skill?

创建和管理基于你们私密聊天数据的朋友数字分身,实现朋友不在线时的温暖陪伴和模拟对话体验。 It is an AI Agent Skill for Claude Code / OpenClaw, with 67 downloads so far.

How do I install friend-skill?

Run "/install friend-skill" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is friend-skill free?

Yes, friend-skill is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does friend-skill support?

friend-skill is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created friend-skill?

It is built and maintained by cp3d1455926-svg (@cp3d1455926-svg); the current version is v1.0.0.

💬 Comments