Claude Code八大隐藏神技：90%用户只会打字聊天

cd your-project
claude remote-control

/remote-control

claude --remote-control "My Project"

# 1. 在BotFather创建bot，获取token
# 2. 安装插件
/plugin install telegram@claude-plugins-official

# 3. 配置token
/telegram:configure <your-bot-token>

# 4. 重启Claude Code并启用channel
claude --channels plugin:telegram@claude-plugins-official

# 5. 在Telegram给bot发消息，获取配对码
# 6. 回到Claude Code完成配对
/telegram:access pair <code>
/telegram:access policy allowlist

flowchart LR
    A[Telegram/Discord<br/>iMessage] -->|推送事件| B[Channel Plugin]
    B -->|注入消息| C[Claude Code会话]
    C -->|分析+操作| D[读取代码/修改文件/运行命令]
    D -->|回复| B
    B -->|发送结果| A

    style A fill:#4A90D9,color:#fff
    style C fill:#D4A44C,color:#fff
    style D fill:#5CB85C,color:#fff

---
name: db-reviewer
description: "Review SQL migrations and database queries. Use when the user writes or modifies SQL, creates migrations, or asks about database performance."
model: haiku
tools:
  - Read
  - Grep
  - Glob
---

You are a senior DBA. Review all SQL for:
- Missing indexes
- N+1 query patterns  
- Data type mismatches
- Migration safety (can it run without locking?)
Always suggest specific improvements with example SQL.

Use the db-reviewer agent to check our recent migrations

{
  "hooks": {
    "PostToolUse": [{
      "matcher": "Edit|Write",
      "hooks": [{
        "type": "command",
        "command": "npx prettier --write ${CLAUDE_FILE_PATH}"
      }]
    }],
    "PreToolUse": [{
      "matcher": "Bash",
      "hooks": [{
        "type": "command",
        "if": "Bash(rm *)",
        "command": "${CLAUDE_PROJECT_DIR}/.claude/hooks/block-rm.sh"
      }]
    }]
  }
}

#!/bin/bash
# .claude/hooks/block-rm.sh
COMMAND=$(jq -r '.tool_input.command')
if echo "$COMMAND" | grep -q 'rm -rf'; then
  jq -n '{hookSpecificOutput: {hookEventName: "PreToolUse", permissionDecision: "deny", permissionDecisionReason: "危险命令已被Hook拦截"}}'
else
  exit 0
fi

mkdir -p ~/.claude/skills/summarize-changes

---
description: "Summarizes uncommitted changes and flags risks. Use when the user asks what changed or wants a commit message."
---

## Current changes
!`git diff HEAD`

## Instructions
Summarize the above changes in 2-3 bullet points.
List risks: missing error handling, hardcoded values, tests needed.
If diff is empty, say "no uncommitted changes".

# 启动后台任务，立即返回
claude --bg "investigate the flaky test in auth module"

# 查看所有后台会话
claude agents --json

# 查看某个后台任务的输出
claude logs <session-id>

# 附加到后台任务（可交互）
claude attach <session-id>

claude --bg --agent my-custom-agent --add-dir ../lib

claude agents          # 查看所有Agent
claude attach <id>     # 附加到Agent
claude logs <id>       # 查看输出
claude stop <id>       # 停止Agent
claude respawn <id>    # 重启Agent（保留对话）
claude rm <id>         # 删除Agent记录

# 分析日志
tail -200 app.log | claude -p "发现异常模式了吗？"

# 审查安全
git diff main --name-only | claude -p "审查这些文件的安全问题"

# 批量翻译
claude -p "translate new strings into French and raise a PR"

claude -p --json-schema '{
  "type": "object",
  "properties": {
    "bugs": {
      "type": "array",
      "items": {
        "type": "object",
        "properties": {
          "file": {"type": "string"},
          "line": {"type": "number"},
          "severity": {"type": "string", "enum": ["critical","warning","info"]},
          "description": {"type": "string"}
        }
      }
    }
  }
}' "审查当前代码库找出潜在bug"

# 最多花5美元
claude -p --max-budget-usd 5.00 "重构auth模块"

# 最多3轮
claude -p --max-turns 3 "修复这个lint错误"

# 裸模式+缓存优化（多人共享场景）
claude --bare --exclude-dynamic-system-prompt-sections -p "查询"

# 快速简单任务
claude --effort low "给这个函数加JSDoc"

# 标准任务
claude --effort medium "修复这个测试"  

# 复杂任务
claude --effort high "重构整个认证模块"

# 最深度思考
claude --effort max "设计新的微服务架构"

claude --permission-mode plan
claude --permission-mode auto

mindmap
  root((Claude Code<br/>隐藏神技))
    远程控制
      手机继续编码
      QR码扫码连接
      Worktree隔离
    外部频道
      Telegram接入
      Discord接入
      iMessage直连
    自定义Agent
      专业化分工
      模型选择省成本
      独立上下文窗口
    Hooks自动化
      文件保存后格式化
      拦截危险命令
      完成后自动通知
    Skills动态注入
      实时git diff嵌入
      可复用命令库
      自然语言触发
    后台并行
      --bg后台任务
      Worktree不冲突
      多Agent协同
    脚本化API
      --bare快速启动
      --json-schema结构化
      管道输入输出
    成本控制
      --effort思考深度
      --max-budget-usd预算锁
      权限模式切换