Agent skill

Claude Code Advanced Guide

by 2025Emma in 2025Emma/vibe-coding-cn

A Chinese-language tutorial on advanced Claude Code use: its core tools, REPL, slash commands, workflows, permissions, memory, MCP, hooks and sub-agents.

MITAuto-check passedAgent Workflows

SKILL.md written in Chinese; this summary is our English description.

Install Claude Code Advanced Guide

skills CLI
$ npx skills add 2025Emma/vibe-coding-cn --skill claude-code-guide -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install 2025Emma/vibe-coding-cn claude-code-guide --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/2025Emma/vibe-coding-cn.git skills-src && mkdir -p .claude/skills && cp -r skills-src/i18n/zh/skills/claude-code-guide .claude/skills/claude-code-guide && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
claude-code-guide
GitHub stars
23k
Used in
1 other repo
Token cost
~1.4k tokens
SKILL.md length
395 words
Files
3 (incl. references)
Skills in repo
13
Repo updated
First seen
Licence
MIT

At a glance

A Chinese-language tutorial on advanced Claude Code use: its core tools, REPL, slash commands, workflows, permissions, memory, MCP, hooks and sub-agents.

  • Works in 3 steps: 文件分析工作流 → 算法验证工作流 → 数据探索工作流
  • Learning Claude Code's core tools and workflows in Chinese
  • SKILL.md covers 何时使用此技能, 快速参考, 核心概念 and MCP 集成, plus 5 more sections
  • Calls npm, pytest and mypy; needs API_KEY

What it does

This is a Chinese-language learning guide for Claude Code, moving from basics to advanced practice. A quick reference lists seven core tools: a JavaScript REPL, Artifacts, Web Search, Web Fetch, Conversation Search, Recent Chats and End Conversation, with short notes on what each does.

Further sections cover a staged workflow for analyzing large files, built-in and custom slash commands defined under .claude/commands, development patterns for file analysis, algorithm verification and data exploration, the tool permission model, project context detection, conversation and experimental persistent memory, MCP server configuration, hooks set in .claude/settings.json, and sub-agent collaboration through the Task tool.

When your agent uses it

  • Learning Claude Code's core tools and workflows in Chinese
  • Setting up MCP servers, hooks or custom slash commands for the first time
  • Looking up a staged approach to analyzing large files with Claude Code

Example prompts

  • “用中文讲解怎么给 Claude Code 配置 MCP 服务器。”
  • “Explain how to create a custom slash command in .claude/commands using this guide.”
  • “Walk me through the large-file analysis workflow from the Claude Code guide.”

Workflow steps

3 steps, taken from the step headings in SKILL.md.

  1. 文件分析工作流
  2. 算法验证工作流
  3. 数据探索工作流

What it can do on your machine

Read from SKILL.md and the folder at commit 9b42dd1. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • npm
    • pytest
    • mypy
    • tsc
    • eslint

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    Links to these hosts (documentation or services it may open):

    • github.com
    • docs.claude.com

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names these keys or tokens, usually read from environment variables:

    • API_KEY

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Claude Code Advanced Guide loads about 1.4k tokens when it runs, and up to ~83k if it reads all its reference files. Until then it costs about 28 tokens; SKILL.md has 395 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~28
When it runs · the whole SKILL.md, loaded when a task matches
~1.4k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~83k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from 2025Emma/vibe-coding-cn at commit 9b42dd1, republished under its MIT licence (© 2025Emma). 395 words, ~1,442 tokens.

Download SKILL.mdSave it as .claude/skills/claude-code-guide/SKILL.md (or your agent's skills folder). This skill also uses 2 other files; get the full folder from GitHub.
name
claude-code-guide
description
Claude Code 高级开发指南 - 全面的中文教程,涵盖工具使用、REPL 环境、开发工作流、MCP 集成、高级模式和最佳实践。适合学习 Claude Code 的高级功能和开发技巧。

Claude Code 高级开发指南

全面的 Claude Code 中文学习指南,涵盖从基础到高级的所有核心概念、工具使用、开发工作流和最佳实践。

何时使用此技能

当需要以下帮助时使用此技能:

  • 学习 Claude Code 的核心功能和工具
  • 掌握 REPL 环境的高级用法
  • 理解开发工作流和任务管理
  • 使用 MCP 集成外部系统
  • 实现高级开发模式
  • 应用 Claude Code 最佳实践
  • 解决常见问题和错误
  • 进行大文件分析和处理

快速参考

Claude Code 核心工具(7个)
  1. REPL - JavaScript 运行时环境

    • 完整的 ES6+ 支持
    • 预加载库:D3.js, MathJS, Lodash, Papaparse, SheetJS
    • 支持 async/await, BigInt, WebAssembly
    • 文件读取:window.fs.readFile()
  2. Artifacts - 可视化输出

    • React, Three.js, 图表库
    • HTML/SVG 渲染
    • 交互式组件
  3. Web Search - 网络搜索

    • 仅美国可用
    • 域名过滤支持
  4. Web Fetch - 获取网页内容

    • HTML 转 Markdown
    • 内容提取和分析
  5. Conversation Search - 对话搜索

    • 搜索历史对话
    • 上下文检索
  6. Recent Chats - 最近对话

    • 访问最近会话
    • 对话历史
  7. End Conversation - 结束对话

    • 清理和总结
    • 会话管理
大文件分析工作流
bash
# 阶段 1:定量评估
wc -l filename.md    # 行数统计
wc -w filename.md    # 词数统计
wc -c filename.md    # 字符数统计

# 阶段 2:结构分析
grep "^#{1,6} " filename.md  # 提取标题层次
grep "```" filename.md       # 识别代码块
grep -c "keyword" filename.md # 关键词频率

# 阶段 3:内容提取
Read filename.md offset=0 limit=50      # 文件开头
Read filename.md offset=N limit=100     # 目标部分
Read filename.md offset=-50 limit=50    # 文件结尾
REPL 高级用法
javascript
// 数据处理
const data = [1, 2, 3, 4, 5];
const sum = data.reduce((a, b) => a + b, 0);

// 使用预加载库
// Lodash
_.chunk([1, 2, 3, 4], 2);  // [[1,2], [3,4]]

// MathJS
math.sqrt(16);  // 4

// D3.js
d3.range(10);  // [0,1,2,3,4,5,6,7,8,9]

// 读取文件
const content = await window.fs.readFile('path/to/file');

// 异步操作
const result = await fetch('https://api.example.com/data');
const json = await result.json();
斜杠命令系统

内置命令:

  • /help - 显示帮助
  • /clear - 清除对话
  • /plugin - 管理插件
  • /settings - 配置设置

自定义命令: 创建 .claude/commands/mycommand.md:

markdown
根据需求执行特定任务的指令

使用:/mycommand

开发工作流模式
1. 文件分析工作流
bash
# 探索 → 理解 → 实现
ls -la                  # 列出文件
Read file.py            # 读取内容
grep "function" file.py # 搜索模式
# 然后实现修改
2. 算法验证工作流
bash
# 设计 → 验证 → 实现
# 1. 在 REPL 中测试逻辑
# 2. 验证边界情况
# 3. 实现到代码
3. 数据探索工作流
bash
# 检查 → 分析 → 可视化
# 1. 读取数据文件
# 2. REPL 中分析
# 3. Artifacts 可视化

核心概念

工具权限系统

自动授予权限的工具:

  • REPL
  • Artifacts
  • Web Search/Fetch
  • Conversation Search

需要授权的工具:

  • Bash (读/写文件系统)
  • Edit (修改文件)
  • Write (创建文件)
项目上下文

Claude 自动识别:

  • Git 仓库状态
  • 编程语言(从文件扩展名)
  • 项目结构
  • 依赖配置
内存系统

对话内存:

  • 存储在当前会话
  • 200K token 窗口
  • 自动上下文管理

持久内存(实验性):

  • 跨会话保存
  • 用户偏好记忆
  • 项目上下文保留

MCP 集成

什么是 MCP?

Model Context Protocol - 连接 Claude 到外部系统的协议。

MCP 服务器配置

配置文件:~/.config/claude/mcp_config.json

json
{
  "mcpServers": {
    "my-server": {
      "command": "node",
      "args": ["path/to/server.js"],
      "env": {
        "API_KEY": "your-key"
      }
    }
  }
}
使用 MCP 工具

Claude 会自动发现 MCP 工具并在对话中使用:

"使用 my-server 工具获取数据"

钩子系统

钩子类型

在 .claude/settings.json 配置:

json
{
  "hooks": {
    "tool-pre-use": "echo 'About to use tool'",
    "tool-post-use": "echo 'Tool used'",
    "user-prompt-submit": "echo 'Processing prompt'"
  }
}
常见钩子用途
  • 自动格式化代码
  • 运行测试
  • Git 提交检查
  • 日志记录
  • 通知发送

高级模式

多代理协作

使用 Task 工具启动子代理:

"启动一个专门的代理来优化这个算法"

子代理特点:

  • 独立上下文
  • 专注单一任务
  • 返回结果到主代理
智能任务管理

使用 TodoWrite 工具:

"创建任务列表来跟踪这个项目"

任务状态:

  • pending - 待处理
  • in_progress - 进行中
  • completed - 已完成
代码生成模式

渐进式开发:

  1. 生成基础结构
  2. 添加核心功能
  3. 实现细节
  4. 测试和优化

验证驱动:

  1. 写测试用例
  2. 实现功能
  3. 运行测试
  4. 修复问题

质量保证

自动化测试
bash
# 运行测试
npm test
pytest

# 类型检查
mypy script.py
tsc --noEmit

# 代码检查
eslint src/
flake8 .
代码审查模式

使用子代理进行审查:

"启动代码审查代理检查这个文件"

审查重点:

  • 代码质量
  • 安全问题
  • 性能优化
  • 最佳实践

错误恢复

Show full SKILL.md (166 more words)Show less
常见错误模式
  1. 工具使用错误

    • 检查权限
    • 验证语法
    • 确认路径
  2. 文件操作错误

    • 确认文件存在
    • 检查读写权限
    • 验证路径正确
  3. API 调用错误

    • 检查网络连接
    • 验证 API 密钥
    • 确认请求格式
渐进式修复策略
  1. 隔离问题
  2. 最小化复现
  3. 逐步修复
  4. 验证解决方案

最佳实践

开发原则
  1. 清晰优先 - 明确需求和目标
  2. 渐进实现 - 分步骤开发
  3. 持续验证 - 频繁测试
  4. 适当抽象 - 合理模块化
工具使用原则
  1. 正确的工具 - 选择合适的工具
  2. 工具组合 - 多工具协同
  3. 权限最小化 - 只请求必要权限
  4. 错误处理 - 优雅处理失败
性能优化
  1. 批量操作 - 合并多个操作
  2. 增量处理 - 处理大文件
  3. 缓存结果 - 避免重复计算
  4. 异步优先 - 使用 async/await

安全考虑

沙箱模型

每个工具在隔离环境中运行:

  • REPL:无文件系统访问
  • Bash:需要明确授权
  • Web:仅特定域名
最佳安全实践
  1. 最小权限 - 仅授予必要权限
  2. 代码审查 - 检查生成的代码
  3. 敏感数据 - 不要共享密钥
  4. 定期审计 - 检查钩子和配置

故障排除

工具无法使用

症状: 工具调用失败

解决方案:

  • 检查权限设置
  • 验证语法正确
  • 确认文件路径
  • 查看错误消息
REPL 性能问题

症状: REPL 执行缓慢

解决方案:

  • 减少数据量
  • 使用流式处理
  • 优化算法
  • 分批处理
MCP 连接失败

症状: MCP 服务器无响应

解决方案:

  • 检查配置文件
  • 验证服务器运行
  • 确认环境变量
  • 查看服务器日志

实用示例

示例 1:数据分析
javascript
// 在 REPL 中
const data = await window.fs.readFile('data.csv');
const parsed = Papa.parse(data, { header: true });
const values = parsed.data.map(row => parseFloat(row.value));
const avg = _.mean(values);
const std = math.std(values);
console.log(`平均值: ${avg}, 标准差: ${std}`);
示例 2:文件搜索
bash
# 在 Bash 中
grep -r "TODO" src/
find . -name "*.py" -type f
示例 3:网络数据获取
"使用 web_fetch 获取 https://api.example.com/data 的内容,
然后在 REPL 中分析 JSON 数据"

参考文件

此技能包含详细文档:

  • README.md (9,594 行) - 完整的 Claude Code 高级指南

包含以下主题:

  • 核心工具深度解析
  • REPL 高级协同模式
  • 开发工作流详解
  • MCP 集成完整指南
  • 钩子系统配置
  • 高级模式和最佳实践
  • 故障排除和安全考虑

使用 view 命令查看参考文件获取详细信息。

资源

注意事项

本指南结合了:

  • 官方功能和公告
  • 实际使用观察到的模式
  • 概念性方法和最佳实践
  • 第三方工具集成

请在使用时参考最新的官方文档。


使用这个技能深入掌握 Claude Code 的强大功能!

© 2025Emma, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 2 other files (references) in i18n/zh/skills/claude-code-guide of 2025Emma/vibe-coding-cn.

  • SKILL.md
  • references/README.md
  • references/index.md

Open the folder on GitHubat commit 9b42dd1

Used in 1 other repository

We found 1 copy of this SKILL.md (exact, near-identical or edited) in other folders, from 1 other GitHub owner. This page covers the copy in 2025Emma/vibe-coding-cn, which our catalogue first saw on October 7, 2026.

Compare with similar skills

Claude Code Advanced Guide next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Claude Code Advanced Guide compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Claude Code Advanced Guide this skill2025Emma/vibe-coding-cn23k1 repos~1.4kAutomated safety check: PassMIT
Claude Automation Recommenderanthropics/claude-plugins-official37k3 repos~2.7kAutomated safety check: NotesApache-2.0
Claude Code Mastery Squadohmyjahh/xquads-squads276—~1.1kAutomated safety check: PassMIT
MCP Integration for Pluginsanthropics/claude-plugins-official37k11 repos~3.1kAutomated safety check: PassApache-2.0
Crush Configurationcharmbracelet/crush29k—~3.7kAutomated safety check: PassCustom licence
OpenpetsOpenPetsHQ/openpets1.3k—~2.1kAutomated safety check: PassMIT

Similar skills

  • Claude Automation Recommender

    anthropics/claude-plugins-official

    Official

    Scans a codebase and suggests which Claude Code hooks, subagents, skills, plugins and MCP servers fit its stack, without changing any files.

    37k GitHub starsUsed in 3 repos~2.7k tokens
    Agent WorkflowsAuto-check: notes
  • Claude Code Mastery Squad

    ohmyjahh/xquads-squads

    Routes a request to one of eight specialist agents covering hooks, skills, subagents, MCP integration and context engineering for Claude Code.

    276 GitHub stars~1.1k tokensUpdated 8 days ago
    Agent WorkflowsAuto-check passed
  • MCP Integration for Plugins

    anthropics/claude-plugins-official

    Official

    Explains how to bundle Model Context Protocol servers in a Claude Code plugin, covering config files, stdio, SSE, HTTP and WebSocket server types, and authentication.

    37k GitHub starsUsed in 11 repos~3.1k tokens
    Agent WorkflowsAuto-check passed
  • Crush Configuration

    charmbracelet/crush

    Explains how to configure the Crush coding agent with crushrc or crush.json, covering providers, models, LSPs, MCP servers, hooks, permissions and config precedence.

    29k GitHub stars~3.7k tokensUpdated today
    Agent WorkflowsAuto-check passed
  • Openpets

    OpenPetsHQ/openpets

    A skill your agent uses whenever the user wants to build, extend, debug, test, validate, locally load, package, or publish an OpenPets plugin; work with the OpenPets Plugin SDK v3, plugin manifest…

    1.3k GitHub stars~2.1k tokensUpdated 10 days ago
    Agent WorkflowsAuto-check passed
  • Mistral Vibe Plugin Creator

    mistralai/mistral-vibe

    Official

    Shows how to build a Vibe plugin package in the Agent Plugins 1.0 format, with a plugin.json manifest and optional skills, MCP servers, hooks and other components.

    5.1k GitHub stars~3.1k tokensUpdated yesterday
    Agent WorkflowsAuto-check passed

More from 2025Emma/vibe-coding-cn

All 13 skills in this repo
  • Claude Cookbooks Reference

    2025Emma/vibe-coding-cn

    Reference of Claude API examples and guides covering tool use, vision, RAG, classification, summarization, text-to-SQL, prompt caching and agent patterns.

    23k GitHub starsUsed in 1 repo~2.2k tokens
    Auto-check passed
  • Proxychains Network Fallback

    2025Emma/vibe-coding-cn

    Retries failed network commands through proxychains4 and a local proxy when it sees timeouts, DNS failures or blocked access.

    23k GitHub starsUsed in 1 repo~1.1k tokens
    Auto-check: notes
  • Skill Authoring and Refactoring

    2025Emma/vibe-coding-cn

    Builds new skills from scattered docs, specs or code, and refactors existing skills for clear triggers, reliable activation and a quality checklist.

    23k GitHub starsUsed in 2 repos~2k tokens
    Auto-check passed
  • CCXT Crypto Exchange Library

    2025Emma/vibe-coding-cn

    Reference help for the CCXT library covering crypto exchange APIs, market data, trading and order management across 150+ exchanges in JavaScript, Python and PHP.

    23k GitHub starsUsed in 2 repos~4.4k tokens
    Auto-check passed
  • CoinGecko API Reference

    2025Emma/vibe-coding-cn

    CoinGecko API documentation - cryptocurrency market data API, price feeds, market cap, volume, historical data. Use when integrating CoinGecko API, building…

    23k GitHub starsUsed in 2 repos~646 tokens
    Auto-check passed
  • Hummingbot Trading Bot Reference

    2025Emma/vibe-coding-cn

    Reference documentation for building, configuring and debugging automated trading strategies, market making and exchange connectors with Hummingbot.

    23k GitHub starsUsed in 1 repo~998 tokens
    Auto-check passed

Questions about Claude Code Advanced Guide

What does Claude Code Advanced Guide do?

A Chinese-language tutorial on advanced Claude Code use: its core tools, REPL, slash commands, workflows, permissions, memory, MCP, hooks and sub-agents. This is a Chinese-language learning guide for Claude Code, moving from basics to advanced practice. A quick reference lists seven core tools: a JavaScript REPL, Artifacts, Web Search, Web Fetch, Conversation Search, Recent Chats and End Conversation, with short notes on what each does.

When should I use Claude Code Advanced Guide?

Claude Code Advanced Guide fits situations like: learning Claude Code's core tools and workflows in Chinese; setting up MCP servers, hooks or custom slash commands for the first time; looking up a staged approach to analyzing large files with Claude Code.

How do I install Claude Code Advanced Guide in Claude Code?

Run `npx skills add 2025Emma/vibe-coding-cn --skill claude-code-guide -a claude-code`. Or copy the skill folder (i18n/zh/skills/claude-code-guide in 2025Emma/vibe-coding-cn) into .claude/skills/claude-code-guide in your project. Claude Code loads it when a task matches its description.

How do I install Claude Code Advanced Guide in Codex?

Run `npx skills add 2025Emma/vibe-coding-cn --skill claude-code-guide -a codex`. Or copy the skill folder (i18n/zh/skills/claude-code-guide in 2025Emma/vibe-coding-cn) into .agents/skills/claude-code-guide in your project. Codex loads it when a task matches its description.

Can I use Claude Code Advanced Guide in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add 2025Emma/vibe-coding-cn --skill claude-code-guide -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/claude-code-guide, .gemini/skills/claude-code-guide, .github/skills/claude-code-guide and .opencode/skills/claude-code-guide in your project.

What does Claude Code Advanced Guide need to run?

Going by SKILL.md and its folder, Claude Code Advanced Guide needs the command-line tools its instructions call (npm, pytest, mypy, tsc and eslint) and credentials named API_KEY.

Does Claude Code Advanced Guide access the network?

SKILL.md names 2 domains. As links in the text: github.com and docs.claude.com. This is read from the text; nothing was executed.

Is Claude Code Advanced Guide safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Claude Code Advanced Guide use?

Claude Code Advanced Guide is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Claude Code Advanced Guide use?

About 1.4k tokens (SKILL.md is roughly 5.8k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 82k tokens, read only when the agent opens those files.

What are the alternatives to Claude Code Advanced Guide?

Skills that share tags, products or a category with Claude Code Advanced Guide: Claude Automation Recommender (anthropics/claude-plugins-official, 37k stars), Claude Code Mastery Squad (ohmyjahh/xquads-squads, 276 stars), MCP Integration for Plugins (anthropics/claude-plugins-official, 37k stars) and Crush Configuration (charmbracelet/crush, 29k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Claude Code Advanced Guide?

2025Emma (a GitHub user) maintains it in 2025Emma/vibe-coding-cn, which has 23,033 GitHub stars. The repository holds 13 skills in this directory. The repository was last updated on December 17, 2025.

Source: 2025Emma/vibe-coding-cn on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.