Token导航 LogoToken导航TokenDH.com
开发敏感数据clawhub未标认证来源可访问clear审计提醒

email-bridge电子邮件桥

Agent Skill

email-bridge 用于整理文档、README、Markdown 和说明材料,适合在 OpenClaw 中需要把零散信息整理成结构清晰的文档时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

4,121

周安装

177

GitHub Stars

公开资料未说明

下载量

1,444
OpenClaw

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

MIT-0

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:email-bridge(电子邮件桥)
来源仓库:https://github.com/ryanchan720/email-bridge
安装命令:
openclaw skills install email-bridge
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 OpenClaw 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

ClawHubOpenClaw
openclaw skills install email-bridge

简介

电子邮件桥技能实时通知、分类邮件并提取验证码。

  • 适用于邮件管理和HTML内容存储,提升信息整理效率。
  • 通过智能分类和提取功能优化邮件处理,适合在OpenClaw中辅助开发任务。
  • 安装命令为openclaw skills install email-bridge,需确认权限和文件操作边界。
  • 建议参考原始README了解具体规则,注意维护状态和网络访问限制。

SKILL.md

name
email-bridge
description
Email management skill for AI assistants with real-time notifications, smart categorization (7 categories), verification code extraction, and HTML content sanitization. Supports Gmail, QQ Mail, and NetEase.
homepage
https://github.com/ryanchan720/email-bridge
source
https://github.com/ryanchan720/email-bridge
version
0.6.3

Email Bridge Skill

Email management skill for OpenClaw. Provides real-time email monitoring with smart categorization and clean notifications for AI assistants.

Features

  • Real-time notifications: IMAP IDLE (QQ/NetEase) + polling (Gmail)
  • Smart categorization: 7 categories with subject-only classification
  • Verification code extraction: Context-aware, low false positive rate
  • HTML content sanitization: Clean text from HTML emails, remove invisible chars
  • Prompt injection protection: Safe email content for AI processing
  • Multi-provider support: Gmail (API), QQ Mail (IMAP), NetEase (IMAP)

Installation

cd skills/email-bridge
pip install -e .

Setup (Manual CLI Required)

⚠️ Security Note: Do NOT share authorization codes in chat. Configure accounts via CLI only.

# Add account (prompts for authorization code securely)
email-bridge accounts add your@qq.com -p qq

# Sync emails
email-bridge sync

# Start daemon for real-time notifications
email-bridge daemon start -d

Getting Authorization Codes

QQ Mail: https://service.mail.qq.com/detail/0/75 (send SMS, get 16-char code)

NetEase (163/126): Settings → POP3/SMTP/IMAP → Enable → Get code

Gmail: Requires OAuth setup (see README.md)

Capabilities

  • Receive emails: Sync and read emails from configured accounts
  • Send emails: Send emails via SMTP
  • Real-time notifications: Push to OpenClaw via openclaw system event
  • Smart categorization: 7 categories with keyword-based classification
  • Verification code extraction: Context-aware extraction with low false positives
  • Link extraction: Extract action links from emails
  • HTML sanitization: Clean text extraction with invisible char removal
  • Prompt injection protection: Sanitize email content for safe AI processing

Email Categories

Subject-only classification for fast, reliable categorization:

CategoryIconDescriptionExample Keywords
verification🔐Verification codes, activation验证码, OTP, activate, 绑定邮箱
security⚠️Security alerts, login warnings安全提醒, security alert, 密码修改
transactional📦Orders, payments, shipping订单确认, receipt, 发货通知
promotion🎁Marketing, promotions, rewards奖励, 优惠, promo, discount
subscription📰Newsletters, digestsnewsletter, 订阅, weekly digest
spam_like🚫Suspected spam中奖, FREE, click here now
normalRegular email(default)

Trigger Keywords

Chinese: 邮箱、邮件、发邮件、查看邮件、验证码、QQ邮箱、Gmail

English: email, mail, send email, check email, verification code

Common Commands

# List recent emails
email-bridge messages list -n 10

# Get verification codes from recent emails
email-bridge codes

# Send email
email-bridge send -a <account_id> -t recipient@example.com -s "Subject" -b "Body"

# Daemon management
email-bridge daemon status
email-bridge daemon stop

Configuration

Configuration file: ~/.email-bridge/config.json

Default configuration (auto-generated, minimal):

{
  "daemon": {
    "poll_interval": 300,
    "notify_openclaw": true
  }
}

Full configuration with all options (customize as needed):

{
  "daemon": {
    "poll_interval": 300,
    "notify_openclaw": true,
    "notification": {
      "include_body": false,
      "body_max_length": 500,
      "include_verification_codes": true,
      "include_links": false
    }
  }
}

Notification Options

OptionDefaultDescription
include_bodyfalseInclude email body preview in notifications
body_max_length500Max characters for body preview
include_verification_codestrueAuto-extract and show verification codes
include_linksfalseInclude action links (verify/reset)

Notifications

When new emails arrive, the daemon sends formatted notifications:

📧 新邮件: account@qq.com

1. 🔐 Google
   您的验证码
   ✨ 验证码: 123456

2. ⚠️ Microsoft
   登录提醒
   📝 We noticed a new sign-in...

3. 🎁 OKX
   150 USDT 奖励等您拿
   📝 亲爱的欧易用户,欧易诚邀您加入邀请好友计划...

HTML Content Processing

HTML-only emails are processed through:

  1. Tag stripping: Remove <style>, <script>, and all HTML tags
  2. Entity decoding: Convert HTML entities to text
  3. Invisible char removal: Remove zero-width spaces, BOM, etc.
  4. Whitespace normalization: Clean up spacing
  5. Prompt injection protection: Remove dangerous patterns

Example: HTML with invisible chars → Clean readable text

Security Features

  • Subject-only classification: No body scanning for privacy
  • Context-aware code extraction: Only extract near verification keywords
  • Invisible char sanitization: Remove U+200B, U+FEFF, U+034F, etc.
  • Prompt injection protection: Filter dangerous instruction patterns
  • Address pattern exclusion: Don't extract numbers from addresses

Data Storage

All data stored locally at ~/.email-bridge/:

~/.email-bridge/
├── email_bridge.db    # SQLite database (accounts, messages)
├── config.json        # Configuration file
├── daemon.pid         # Daemon process ID
├── daemon.log         # Logs
└── gmail/
    ├── credentials.json  # OAuth credentials
    └── token_*.json      # OAuth tokens

⚠️ Credentials are stored unencrypted. Protect this directory.

Revoking Access

# Stop daemon
email-bridge daemon stop

# Remove all stored data
rm -rf ~/.email-bridge

# For Gmail: revoke at https://myaccount.google.com/permissions
# For QQ/NetEase: regenerate authorization codes in email settings

Dependencies

All from PyPI:

  • click >= 8.0
  • pydantic >= 2.0
  • imaplib2 >= 3.6
  • google-api-python-client >= 2.0 (Gmail only)
  • google-auth-oauthlib >= 1.0 (Gmail only)

Security Notes

  1. Never share authorization codes in chat - use CLI interactively
  2. Credentials stored unencrypted - protect ~/.email-bridge/ directory
  3. Email content is sanitized - prompt injection protection enabled
  4. Daemon runs with user privileges - no elevated access needed
  5. Subject-only classification - privacy-conscious processing

Changelog

v0.6.2

  • Add PROMOTION category for marketing emails (🎁 icon)
  • Add TRANSACTIONAL category for orders/shipping (📦 icon)
  • Expand keyword pools for all categories
  • Add invisible character sanitization (U+200B, U+FEFF, U+034F, etc.)
  • Improve HTML-to-text extraction
  • Update documentation (DESIGN.md, README.md)

v0.6.1

  • Add IDLE keepalive (60s timeout) for connection stability
  • Add sync retry mechanism (up to 3 retries)
  • Improve daemon reliability for flaky networks

v0.6.0

  • Smart notification format based on email category
  • Prompt injection protection with sanitize_for_notification()
  • HTML-to-text fallback for HTML-only emails
  • Subject-only classification for privacy
  • Context-aware verification code extraction
  • Category icons (🔐 ⚠️ 📦 🎁 📰 🚫)

v0.5.7

  • Initial ClawHub release
  • Gmail, QQ Mail, NetEase support
  • IMAP IDLE real-time notifications
  • Verification code extraction
  • Link extraction

适合场景

01

OpenClaw 用户查找和安装 Skill 时

02

用户想查找某类 Agent Skill 时

03

需要根据任务场景推荐可安装能力包时

04

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

补充不同宿主或平台的使用分布数据

能力 5

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

OpenClaw

81.02%
按下载量换算1,170

安全审计

VirusTotal

可疑

ClawScan

通过

Static analysis

通过

权限和风险

敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills