Token导航 LogoToken导航TokenDH.com
开发操作浏览器clawhub未标认证来源可访问clear审计提醒

oauth-helperOAuth 助手

Agent Skill

用于辅助安全审计、权限检查、凭据风险、认证流程和常见漏洞排查。它适合让 Agent 梳理敏感配置、检查依赖风险、分析鉴权逻辑或生成安全复核清单。使用时不能把工具输出直接当最终结论,涉及密钥、令牌、用户数据或生产系统时,应先确认最小权限、脱敏方式和操作边界。

总安装

91,217

周安装

3,690

GitHub Stars

3

下载量

28,634
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:oauth-helper(OAuth 助手)
来源仓库:https://github.com/helloliuyongsheng-bot/oauth-helper
安装命令:
openclaw skills install oauth-helper
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 oauth-helper

简介

通过 Telegram 用户确认,自动化 OAuth 登录流程。

  • 支持 7 个提供商:Google、Apple、Microsoft、GitHub、Discord、微信、QQ。
  • 特点:
  • - 自动检测登录页面上可用的 OAuth 选项
  • - 当存在多个选项时,要求用户通过 Telegram 进行选择
  • - 授权前确认
  • - 自动处理帐户选择和同意页面

SKILL.md

name
oauth-helper
description
|
Features

OAuth Helper

Automate OAuth login with Telegram confirmation. Supports 7 major providers.

Supported Providers

ProviderStatusDetection Domain
Googleaccounts.google.com
Appleappleid.apple.com
Microsoftlogin.microsoftonline.com, login.live.com
GitHubgithub.com/login/oauth
Discorddiscord.com/oauth2
WeChatopen.weixin.qq.com
QQgraph.qq.com

Prerequisites

  1. Clawd browser logged into the OAuth providers (one-time setup)
  2. Telegram channel configured

Core Workflow

Flow A: Login Page with Multiple OAuth Options

When user requests to login to a website:

1. Open website login page
2. Scan page for available OAuth buttons
3. Send Telegram message:
   "🔐 [Site] supports these login methods:
    1️⃣ Google
    2️⃣ Apple  
    3️⃣ GitHub
    Reply with number to choose"
4. Wait for user reply (60s timeout)
5. Click the selected OAuth button
6. Enter Flow B

Flow B: OAuth Authorization Page

When on an OAuth provider's page:

1. Detect OAuth page type (by URL)
2. Extract target site info
3. Send Telegram: "🔐 [Site] requests [Provider] login. Confirm? Reply yes"
4. Wait for "yes" (60s timeout)
5. Execute provider-specific click sequence
6. Wait for redirect back to original site
7. Send: "✅ Login successful!"

Detection Patterns

Google

URL patterns:
- accounts.google.com/o/oauth2
- accounts.google.com/signin/oauth
- accounts.google.com/v3/signin

Apple

URL patterns:
- appleid.apple.com/auth/authorize
- appleid.apple.com/auth/oauth2

Microsoft

URL patterns:
- login.microsoftonline.com/common/oauth2
- login.microsoftonline.com/consumers
- login.live.com/oauth20

GitHub

URL patterns:
- github.com/login/oauth/authorize
- github.com/login
- github.com/sessions/two-factor

Discord

URL patterns:
- discord.com/oauth2/authorize
- discord.com/login
- discord.com/api/oauth2

WeChat

URL patterns:
- open.weixin.qq.com/connect/qrconnect
- open.weixin.qq.com/connect/oauth2

QQ

URL patterns:
- graph.qq.com/oauth2.0/authorize
- ssl.xui.ptlogin2.qq.com
- ui.ptlogin2.qq.com

Click Sequences by Provider

Google

Account selector: [data-identifier], .JDAKTe
Auth buttons: button:has-text("Allow"), button:has-text("Continue")

Apple

Email input: input[type="email"], #account_name_text_field
Password: input[type="password"], #password_text_field  
Continue: button#sign-in, button:has-text("Continue")
Trust device: button:has-text("Trust")

Microsoft

Account selector: .table-row[data-test-id]
Email input: input[name="loginfmt"]
Password: input[name="passwd"]
Next: button#idSIButton9
Accept: button#idBtn_Accept

GitHub

Email: input#login_field
Password: input#password
Sign in: input[type="submit"]
Authorize: button[name="authorize"]
2FA: input#app_totp

Discord

Email: input[name="email"]
Password: input[name="password"]
Login: button[type="submit"]
Authorize: button:has-text("Authorize")

WeChat

Method: QR code scan
- Screenshot QR code to user
- Wait for mobile scan confirmation
- Detect page redirect

QQ

Method: QR code or password login
QR: Screenshot to user
Password mode:
  - Switch: a:has-text("密码登录")
  - Username: input#u
  - Password: input#p
  - Login: input#login_button

OAuth Button Detection

Scan login pages for these selectors:

ProviderSelectorsCommon Text
Google[data-provider="google"], .google-btn"Continue with Google"
Apple[data-provider="apple"], .apple-btn"Sign in with Apple"
Microsoft[data-provider="microsoft"]"Sign in with Microsoft"
GitHub[data-provider="github"]"Continue with GitHub"
Discord[data-provider="discord"]"Login with Discord"
WeChat.wechat-btn, img[src*="wechat"]"WeChat Login"
QQ.qq-btn, img[src*="qq"]"QQ Login"

One-Time Setup

Login to each provider in clawd browser:

# Google
browser action=navigate profile=clawd url=https://accounts.google.com

# Apple
browser action=navigate profile=clawd url=https://appleid.apple.com

# Microsoft  
browser action=navigate profile=clawd url=https://login.live.com

# GitHub
browser action=navigate profile=clawd url=https://github.com/login

# Discord
browser action=navigate profile=clawd url=https://discord.com/login

# WeChat/QQ - Use QR scan, no pre-login needed

Error Handling

  • No "yes" reply → Cancel and notify user
  • 2FA required → Prompt user to enter code manually
  • QR timeout → Re-screenshot new QR code
  • Login failed → Screenshot and send to user for debugging

Usage Example

User: Login to Kaggle for me

Agent:
1. Navigate to kaggle.com/account/login
2. Detect Google/Facebook/Yahoo options
3. Send: "🔐 Kaggle supports:
   1️⃣ Google
   2️⃣ Facebook
   3️⃣ Yahoo
   Reply number to choose"
4. User replies: 1
5. Click Google login
6. Detect Google OAuth page
7. Send: "🔐 Kaggle requests Google login. Confirm? Reply yes"
8. User replies: yes
9. Select account, click Continue
10. Send: "✅ Logged into Kaggle!"

Version History

  • v1.0.0 - Initial release with 7 OAuth providers

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

73.69%
按下载量换算21,100

安全审计

VirusTotal

通过

ClawScan

可疑

Static analysis

未展示

权限和风险

操作浏览器

该 Skill 可能涉及浏览器控制能力,使用时可能读取或操作网页内容,需要在受控环境中确认权限边界。

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