Skip to content
Back to search

ai-partner-chat

🌐 Language English | 简体中文 | 繁體中文 | 日本語 | 한국어 | हिन्दी | ไทย | Français | Deutsch | Español | Italiano | Русский | Português | Nederlands | Polski | العربية | فارسی | Türkçe | Tiếng Việt | Bahasa…

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/eze-is--ai-partner-chat/badge.svg)](/listings/eze-is--ai-partner-chat)

HTML

<a href="/listings/eze-is--ai-partner-chat"><img src="/api/listings/eze-is--ai-partner-chat/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

SKILL.md · Manifest · 368 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.

  1. ---
  2. name: ai-partner-chat
  3. description: 基于用户画像和向量化笔记提供个性化对话。当用户需要个性化交流、上下文感知的回应,或希望 AI 记住并引用其之前的想法和笔记时使用。
  4. ---
  5. # AI Partner Chat
  6. ## Overview
  7. Provide personalized, context-aware conversations by integrating user persona, AI persona, and vectorized personal notes. This skill enables AI to remember and reference the user's previous thoughts, preferences, and knowledge base, creating a more coherent and personalized interaction experience.
  8. ## Prerequisites
  9. Before first use, complete these steps in order:
  10. 1. **Create directory structure**
  11. ```bash
  12. mkdir -p config notes vector_db scripts
  13. ```
  14. 2. **Set up Python environment**
  15. ```bash
  16. python3 -m venv venv
  17. ./venv/bin/pip install -r .claude/skills/ai-partner-chat/scripts/requirements.txt
  18. ```
  19. Note: First run will download embedding model (~4.3GB)
  20. 3. **Generate persona templates**
  21. Copy from `.claude/skills/ai-partner-chat/assets/` to `config/`:
  22. - `user-persona-template.md` → `config/user-persona.md`
  23. - `ai-persona-template.md` → `config/ai-persona.md`
  24. 4. **User adds notes**
  25. Place markdown notes in `notes/` directory (any format/structure)
  26. 5. **Initialize vector database** (see section 1.2 below)
  27. Now proceed to Core Workflow →
  28. ## Core Workflow
  29. ### 1. Initial Setup
  30. Before using this skill for the first time, complete the following setup:
  31. #### 1.1 Create Persona Files
  32. Create two Markdown files to define interaction parameters:
  33. **User Persona** (`user-persona.md`):
  34. - Define user's background, expertise, interests
  35. - Specify communication preferences and working style
  36. - Include learning goals and current projects
  37. - Use template: `assets/user-persona-template.md`
  38. **AI Persona** (`ai-persona.md`):
  39. - Define AI's role and expertise areas
  40. - Specify communication style and tone
  41. - Set interaction guidelines and response strategies
  42. - Define how to use user context and reference notes
  43. - Use template: `assets/ai-persona-template.md`
  44. #### 1.2 Initialize Vector Database
  45. This skill uses **AI Agent approach** for intelligent note chunking:
  46. **When you initialize the vector database, Claude Code will:**
  47. 1. Read notes from `<project_root>/notes/` directory
  48. 2. **Analyze each note's format** (daily logs, structured docs, continuous text, etc.)
  49. 3. **Generate custom chunking code** tailored to that specific note
  50. 4. Execute the code to produce chunks conforming to `chunk_schema.Chunk` format
  51. 5. Generate embeddings using **BAAI/bge-m3** (optimized for Chinese text)
  52. 6. Store in ChromaDB at `<project_root>/vector_db/`
  53. **Key advantages:**
  54. - ✅ No pre-written chunking strategies needed
  55. - ✅ Each note gets optimal chunking based on its actual structure
  56. - ✅ True AI Agent - generates tools on demand, not calling pre-built tools
  57. **Chunk Format Requirement:**
  58. All chunks must conform to this schema (see `scripts/chunk_schema.py`):
  59. ```python
  60. {
  61. 'content': 'chunk text content',
  62. 'metadata': {
  63. 'filename': 'note.md', # Required
  64. 'filepath': '/path/to/file', # Required
  65. 'chunk_id': 0, # Required
  66. 'chunk_type': 'date_entry', # Required
  67. 'date': '2025-11-07', # Optional
  68. 'title': 'Section title', # Optional
  69. }
  70. }
  71. ```
  72. #### Implementation Requirements
  73. **Location**: Create `<project_root>/scripts/chunk_and_index.py`
  74. **Required structure**:
  75. ```python
  76. # Import provided utilities
  77. import sys
  78. from pathlib import Path
  79. sys.path.insert(0, str(Path(__file__).parent.parent / ".claude/skills/ai-partner-chat/scripts"))
  80. from chunk_schema import Chunk, validate_chunk
  81. from vector_indexer import VectorIndexer
  82. def chunk_note_file(filepath: str) -> List[Chunk]:
  83. """
  84. Analyze THIS file's format and generate appropriate chunks.
  85. Each chunk must conform to chunk_schema.Chunk format:
  86. {
  87. 'content': 'text',
  88. 'metadata': {
  89. 'filename': 'file.md',
  90. 'filepath': '/path/to/file',
  91. 'chunk_id': 0,
  92. 'chunk_type': 'your_label'
  93. }
  94. }
  95. """
  96. # TODO: Analyze actual file format (NOT template-based)
  97. # TODO: Generate chunks based on analysis
  98. # TODO: Validate each chunk with validate_chunk()
  99. pass
  100. def main():
  101. # Initialize vector database
  102. indexer = VectorIndexer(db_path="./vector_db")
  103. indexer.initialize_db()
  104. # Process all note files
  105. all_chunks = []
  106. for note_file in Path("./notes").glob("**/*"):
  107. if note_file.is_file():
  108. chunks = chunk_note_file(str(note_file))
  109. all_chunks.extend(chunks)
  110. # Index chunks
  111. indexer.index_chunks(all_chunks)
  112. if __name__ == "__main__":
  113. main()
  114. ```
  115. **Execute**: `./venv/bin/python scripts/chunk_and_index.py`
  116. **Key points:**
  117. - The `chunk_note_file()` function logic should be **dynamically created** based on analyzing actual file content
  118. - Do NOT copy chunking strategies from examples or templates
  119. - Each file may have different format - analyze individually
  120. - Only requirement: output must conform to `chunk_schema.Chunk`
  121. ### 2. Conversation Workflow
  122. For each user query, follow this process:
  123. #### 2.1 Load Personas
  124. Read both persona files to understand:
  125. - User's background, preferences, and communication style
  126. - AI's role definition and interaction guidelines
  127. - How to appropriately reference context
  128. #### 2.2 Retrieve Relevant Notes
  129. Query the vector database to find the top 5 most semantically similar notes:
  130. ```python
  131. from scripts.vector_utils import get_relevant_notes
  132. # Query for relevant context
  133. relevant_notes = get_relevant_notes(
  134. query=user_query,
  135. db_path="./vector_db",
  136. top_k=5
  137. )
  138. ```
  139. Or use the command-line tool:
  140. ```bash
  141. python scripts/query_notes.py "user query text" --top-k 5
  142. ```
  143. #### 2.3 Construct Context
  144. Combine the following elements to inform the response:
  145. 1. **User Persona**: Background, preferences, expertise
  146. 2. **AI Persona**: Role, communication style, guidelines
  147. 3. **Relevant Notes** (top 5): User's previous thoughts and knowledge
  148. 4. **Current Conversation**: Ongoing chat history
  149. #### 2.4 Generate Response
  150. Synthesize a response that:
  151. - Aligns with both persona definitions
  152. - Naturally references relevant notes when applicable
  153. - Maintains continuity with user's knowledge base
  154. - Follows the AI persona's communication guidelines
  155. **When Referencing Notes:**
  156. - Use natural phrasing: "Based on your previous note about..."
  157. - Make connections: "This relates to what you mentioned in..."
  158. - Avoid robotic citations: integrate context smoothly
  159. **Example Response Pattern:**
  160. ```
  161. [Acknowledge user's query in preferred communication style]
  162. [Incorporate relevant note context naturally if applicable]
  163. "I remember you mentioned [insight from note] - this connects well with..."
  164. [Provide main response following AI persona guidelines]
  165. [Optional: Ask follow-up question based on user's learning style]
  166. ```
  167. ### 3. Maintenance
  168. #### Adding New Notes
  169. When the user creates new notes, add them to the vector database:
  170. ```bash
  171. python scripts/add_note.py /path/to/new_note.md
  172. ```
  173. #### Updating Personas
  174. Personas can be updated anytime by editing the Markdown files. Changes take effect in the next conversation.
  175. #### Reinitializing Database
  176. To completely rebuild the vector database:
  177. ```bash
  178. python scripts/init_vector_db.py /path/to/notes --db-path ./vector_db
  179. ```
  180. This will delete the existing database and re-index all notes.
  181. ## Technical Details
  182. ### Data Architecture
  183. **User data is stored in project root**, not inside the skill directory:
  184. ```
  185. <project_root>/
  186. ├── notes/ # User's markdown notes
  187. ├── vector_db/ # ChromaDB vector database
  188. ├── venv/ # Python dependencies
  189. ├── config/
  190. │ ├── user-persona.md # User persona definition
  191. │ └── ai-persona.md # AI persona definition
  192. └── .claude/skills/ai-partner-chat/ # Skill code (can be deleted/reinstalled)
  193. ├── SKILL.md
  194. └── scripts/
  195. ├── chunk_schema.py # Chunk format specification
  196. ├── vector_indexer.py # Core indexing utilities
  197. └── vector_utils.py # Query utilities
  198. ```
  199. **Design principles:**
  200. - ✅ User data (notes, personas, vectors) lives in project root
  201. - ✅ Easy to backup, migrate, or share across skills
  202. - ✅ Skill code is stateless and replaceable
  203. ### AI Agent Chunking
  204. **Philosophy**: Instead of pre-written chunking strategies, Claude Code analyzes each note and generates optimal chunking code on the fly.
  205. **How it works:**
  206. 1. Claude reads a note file
  207. 2. Analyzes format features (date headers, section titles, separators, etc.)
  208. 3. Writes Python code that chunks this specific note optimally
  209. 4. Executes the code to produce chunks
  210. 5. Validates chunks against `chunk_schema.Chunk` format
  211. 6. Indexes chunks using `vector_indexer.py`
  212. **Benefits:**
  213. - Adapts to any note format without pre-programming
  214. - Can handle mixed formats, unusual structures, or evolving note styles
  215. - True "vibe coding" approach - tools are created when needed
  216. ### Vector Database
  217. - **Storage**: ChromaDB (persistent local storage at `<project_root>/vector_db/`)
  218. - **Embedding Model**: BAAI/bge-m3 (multilingual, optimized for Chinese)
  219. - **Similarity Metric**: Cosine similarity
  220. - **Chunking**: AI-generated custom code per note
  221. ### Scripts
  222. - `chunk_schema.py`: Defines required chunk format specification
  223. - `vector_indexer.py`: Core utilities for embedding generation and ChromaDB indexing
  224. - `vector_utils.py`: Query utilities for retrieving relevant chunks
  225. - `requirements.txt`: Python dependencies (chromadb, sentence-transformers)
  226. **Note**: No pre-written chunking scripts. Chunking is done by Claude Code dynamically.
  227. ### File Structure
  228. ```
  229. <project_root>/
  230. ├── notes/ # User's notes (managed by user)
  231. │ └── *.md
  232. ├── vector_db/ # Vector database (auto-generated)
  233. ├── venv/ # Python environment
  234. ├── config/ # User configuration
  235. │ ├── user-persona.md
  236. │ └── ai-persona.md
  237. └── .claude/skills/ai-partner-chat/
  238. ├── SKILL.md # This file
  239. ├── scripts/
  240. │ ├── chunk_schema.py # Chunk format spec
  241. │ ├── vector_indexer.py # Indexing utilities
  242. │ ├── vector_utils.py # Query utilities
  243. │ └── requirements.txt # Dependencies
  244. └── assets/
  245. ├── user-persona-template.md
  246. └── ai-persona-template.md
  247. ```
  248. ## Best Practices
  249. ### Persona Design
  250. - **Be Specific**: Vague personas lead to generic responses
  251. - **Include Examples**: Show desired interaction patterns in AI persona
  252. - **Update Regularly**: Refine personas based on conversation quality
  253. - **Balance Detail**: Provide enough context without overwhelming
  254. ### Note Management
  255. - **Any Format Welcome**: AI Agent approach adapts to your note structure
  256. - **Meaningful Content**: Rich, substantive notes yield better retrieval
  257. - **Regular Updates**: Add new notes to `<project_root>/notes/` anytime
  258. - **Rebuild When Needed**: Re-index when note collection changes significantly
  259. ### Context Integration
  260. - **Natural References**: Avoid forced citations - only reference when genuinely relevant
  261. - **Connection Quality**: Prioritize meaningful connections over quantity
  262. - **Respect Privacy**: Be mindful of sensitive information in notes
  263. - **Conversation Flow**: Don't let note references disrupt natural dialogue
  264. ## Troubleshooting
  265. **Database Connection Errors:**
  266. - Ensure `<project_root>/vector_db/` directory exists and is writable
  267. - Check that Python dependencies are installed in venv
  268. **Poor Retrieval Quality:**
  269. - Try re-indexing with Claude Code analyzing notes fresh
  270. - Verify notes contain substantial content (not just titles)
  271. - Consider increasing `top_k` value for more context
  272. **Chunking Issues:**
  273. - If chunks are too large/small, ask Claude to adjust chunking strategy
  274. - Review generated chunking code and provide feedback
  275. - Ensure notes have clear structure for better chunking
README.md · README · 242 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.

  1. <div align="right">
  2. <details>
  3. <summary >🌐 Language</summary>
  4. <div>
  5. <div align="center">
  6. <a href="https://openaitx.github.io/view.html?user=eze-is&project=ai-partner-chat&lang=en">English</a>
  7. | <a href="https://openaitx.github.io/view.html?user=eze-is&project=ai-partner-chat&lang=zh-CN">简体中文</a>
  8. | <a href="https://openaitx.github.io/view.html?user=eze-is&project=ai-partner-chat&lang=zh-TW">繁體中文</a>
  9. | <a href="https://openaitx.github.io/view.html?user=eze-is&project=ai-partner-chat&lang=ja">日本語</a>
  10. | <a href="https://openaitx.github.io/view.html?user=eze-is&project=ai-partner-chat&lang=ko">한국어</a>
  11. | <a href="https://openaitx.github.io/view.html?user=eze-is&project=ai-partner-chat&lang=hi">हिन्दी</a>
  12. | <a href="https://openaitx.github.io/view.html?user=eze-is&project=ai-partner-chat&lang=th">ไทย</a>
  13. | <a href="https://openaitx.github.io/view.html?user=eze-is&project=ai-partner-chat&lang=fr">Français</a>
  14. | <a href="https://openaitx.github.io/view.html?user=eze-is&project=ai-partner-chat&lang=de">Deutsch</a>
  15. | <a href="https://openaitx.github.io/view.html?user=eze-is&project=ai-partner-chat&lang=es">Español</a>
  16. | <a href="https://openaitx.github.io/view.html?user=eze-is&project=ai-partner-chat&lang=it">Italiano</a>
  17. | <a href="https://openaitx.github.io/view.html?user=eze-is&project=ai-partner-chat&lang=ru">Русский</a>
  18. | <a href="https://openaitx.github.io/view.html?user=eze-is&project=ai-partner-chat&lang=pt">Português</a>
  19. | <a href="https://openaitx.github.io/view.html?user=eze-is&project=ai-partner-chat&lang=nl">Nederlands</a>
  20. | <a href="https://openaitx.github.io/view.html?user=eze-is&project=ai-partner-chat&lang=pl">Polski</a>
  21. | <a href="https://openaitx.github.io/view.html?user=eze-is&project=ai-partner-chat&lang=ar">العربية</a>
  22. | <a href="https://openaitx.github.io/view.html?user=eze-is&project=ai-partner-chat&lang=fa">فارسی</a>
  23. | <a href="https://openaitx.github.io/view.html?user=eze-is&project=ai-partner-chat&lang=tr">Türkçe</a>
  24. | <a href="https://openaitx.github.io/view.html?user=eze-is&project=ai-partner-chat&lang=vi">Tiếng Việt</a>
  25. | <a href="https://openaitx.github.io/view.html?user=eze-is&project=ai-partner-chat&lang=id">Bahasa Indonesia</a>
  26. | <a href="https://openaitx.github.io/view.html?user=eze-is&project=ai-partner-chat&lang=as">অসমীয়া</
  27. </div>
  28. </div>
  29. </details>
  30. </div>
  31. # AI Partner Chat
  32. 一个 Claude Skills 项目,让 AI 成为你的个性化对话伙伴。
  33. ## 项目简介
  34. AI Partner Chat 通过整合用户画像、AI 画像和向量化的个人笔记,提供个性化的、上下文感知的对话体验。这个技能让 AI 能够记住并引用你之前的想法、偏好和知识库,创造出更加连贯和个性化的交互体验。
  35. ## 核心功能
  36. ### 🎭 双画像系统
  37. - **用户画像**:定义你的背景、专长、兴趣和沟通偏好
  38. - **AI 画像**:定制 AI 的角色定位、沟通风格和交互方式
  39. ### 📝 智能笔记检索
  40. - 自动索引你的 Markdown 笔记
  41. - 根据对话内容智能检索相关历史记录
  42. - 在对话中自然引用你的过往想法和笔记
  43. ### 💬 个性化对话
  44. - 基于你的画像和笔记生成个性化回应
  45. - 保持对话的上下文连贯性
  46. - 像朋友一样自然地引用你的想法,而非机械地"根据记录"
  47. ## 使用场景
  48. 当你需要:
  49. - 个性化交流,而非通用回复
  50. - 上下文感知的回应,AI 能记住你的背景
  51. - AI 记住并引用你之前的想法和笔记
  52. - 持续性的对话体验,而非每次都是全新开始
  53. ## 安装与使用
  54. ### 安装技能
  55. 将此项目复制到你的工作目录下的 `.claude/skills/` 文件夹中:
  56. ```
  57. <你的项目根目录>/
  58. └── .claude/
  59. └── skills/
  60. └── ai-partner-chat/ # 本技能包
  61. ├── assets/
  62. ├── scripts/
  63. ├── SKILL.md
  64. └── README.md
  65. ```
  66. ### 使用技能
  67. 在 Claude Code 中,发送以下指令即可启用此技能:
  68. ```
  69. 遵循 ai-partner-chat 对话
  70. ```
  71. AI agent 会自动:
  72. - 读取技能配置和指引
  73. - 创建必要的目录结构(`notes/`、`config/`、`vector_db/` 等)
  74. - 根据你的需求进行初始化
  75. ### 初始化流程
  76. #### 方式一:让 AI 自动创建并配置
  77. 首次使用时,直接告诉 AI:
  78. ```
  79. 我刚刚在 notes 里放入了对应的笔记,请根据笔记内容,进行向量化;并基于笔记内容,推测并更新 user-persona.md,以及最适合我的 ai-persona.md
  80. ```
  81. AI agent 会:
  82. 1. 分析 `notes/` 目录中的笔记内容
  83. 2. 根据笔记格式智能分块并创建向量数据库
  84. 3. 基于笔记内容推测你的背景和偏好
  85. 4. 自动生成并更新 `config/user-persona.md`
  86. 5. 根据你的特点推荐并创建 `config/ai-persona.md`
  87. #### 方式二:手动配置画像
  88. 如果你想自己定义画像:
  89. 1. AI agent 会自动从模板创建画像文件到 `config/` 目录
  90. 2. 你可以手动编辑这些文件来定制画像
  91. 3. 然后告诉 AI 进行向量化处理
  92. ### 开始对话
  93. 配置完成后,每次使用只需发送:
  94. ```
  95. 遵循 ai-partner-chat 对话
  96. ```
  97. AI 将:
  98. - 读取你的画像了解你的背景
  99. - 检索相关的历史笔记
  100. - 生成个性化的、上下文感知的回应
  101. ## 项目结构
  102. ### 技能包结构(位于 `.claude/skills/ai-partner-chat/`)
  103. ```
  104. ai-partner-chat/
  105. ├── assets/ # 画像模板
  106. │ ├── user-persona-template.md
  107. │ └── ai-persona-template.md
  108. ├── scripts/ # 核心脚本
  109. │ ├── chunk_schema.py
  110. │ ├── vector_indexer.py
  111. │ ├── vector_utils.py
  112. │ └── requirements.txt
  113. ├── SKILL.md # 技能详细文档(AI agent 会读取此文件)
  114. └── README.md # 本文件
  115. ```
  116. ### 用户数据目录(位于项目根目录)
  117. AI agent 会在你的项目根目录创建以下结构:
  118. ```
  119. <项目根目录>/
  120. ├── notes/ # 你的笔记(由你或 AI agent 创建)
  121. ├── config/ # 画像配置(由 AI agent 创建)
  122. │ ├── user-persona.md
  123. │ └── ai-persona.md
  124. ├── vector_db/ # 向量数据库(由 AI agent 创建)
  125. └── venv/ # Python 虚拟环境(由 AI agent 创建)
  126. ```
  127. **重要**:用户数据与技能包分离,便于备份和迁移。
  128. ## 工作流程
  129. 1. **加载画像**:读取用户画像和 AI 画像,理解交互背景
  130. 2. **检索笔记**:根据用户查询,从向量数据库中检索最相关的笔记
  131. 3. **构建上下文**:整合画像信息、相关笔记和对话历史
  132. 4. **生成回应**:基于上下文生成个性化、自然的回应
  133. ## 特色亮点
  134. ### 🤖 AI Agent 智能分块
  135. 系统会分析每篇笔记的实际格式,动态生成最适合的分块策略,而非使用预设模板。这意味着无论你的笔记是什么格式,都能得到最优处理。
  136. ### 🎯 自然引用
  137. AI 会像回忆一样自然地引入你的过往信息,不会生硬地说"根据记录",而是流畅地融入对话中。
  138. ### 📦 数据独立
  139. 你的所有数据(笔记、画像、向量库)都存储在项目根目录,易于备份、迁移或在不同技能间共享。
  140. ## 最佳实践
  141. ### 画像设计
  142. - **具体明确**:模糊的画像会导致通用回复
  143. - **包含示例**:在 AI 画像中展示期望的交互模式
  144. - **定期更新**:根据对话质量不断优化画像
  145. ### 笔记管理
  146. - **格式自由**:系统能适应任何笔记结构
  147. - **内容丰富**:有深度的笔记能带来更好的检索效果
  148. - **及时更新**:新笔记记得添加到索引中
  149. ### 对话体验
  150. - **自然引用**:只在真正相关时才引用笔记
  151. - **保持流畅**:不要让引用打断对话的自然节奏
  152. - **关注质量**:优先考虑有意义的连接,而非数量
  153. ## 维护与更新
  154. ### 添加新笔记
  155. 将新笔记放入 `notes/` 目录后,告诉 AI:
  156. ```
  157. 我刚刚在 notes 里添加了新笔记,请更新向量数据库
  158. ```
  159. AI agent 会自动分析新笔记并更新索引。
  160. ### 更新画像
  161. 你可以直接编辑 `config/` 目录下的画像文件,或者告诉 AI:
  162. ```
  163. 请根据我最近的笔记内容,更新 user-persona.md 和 ai-persona.md
  164. ```
  165. ### 重建索引
  166. 当笔记结构发生重大变化时,告诉 AI:
  167. ```
  168. 请重新初始化向量数据库
  169. ```
  170. AI agent 会重新分析所有笔记并重建索引。
  171. ## 注意事项
  172. - **首次运行**:AI agent 首次创建向量数据库时会自动下载嵌入模型(约 4.3GB),请耐心等待
  173. - **Python 环境**:AI agent 会自动创建虚拟环境并安装所需依赖
  174. - **数据存储**:所有数据(笔记、画像、向量库)存储在项目根目录,而非技能包目录内
  175. - **技能位置**:确保技能包位于 `.claude/skills/ai-partner-chat/` 目录下
  176. ## 更多信息
  177. 详细的技术文档和使用说明请参考 `SKILL.md` 文件。
  178. ---
  179. 让 AI 成为真正了解你的对话伙伴,而不仅仅是一个工具。

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.