← 返回 Skills 市场
Email Analyzer
作者
farmerboybi-bot
· GitHub ↗
· v1.0.0
· MIT-0
224
总下载
0
收藏
1
当前安装
1
版本数
在 OpenClaw 中安装
/install email-analyzer
功能描述
分析并清理 [email protected] 邮箱中的邮件，按固化关键词分类后生成报告，待确认后执行安全删除操作。
使用说明 (SKILL.md)
📧 Email Analyzer 技能（固化版）
创建时间：2026-03-06
固化状态：🔒 锁死 - 改动需 Wood 哥书面同意
执行者：湫儿
监督人：Wood 哥
📋 技能说明
用途：分析、清理 [email protected] 邮箱邮件
唯一性：这是访问 Wood 哥邮箱的唯一合法方式
🔒 固化配置（禁止修改）
| 配置项 | 值 | 备注 |
|---|---|---|
| IMAP 服务器 | imap.126.com:993 |
126 邮箱 |
| 邮箱账户 | [email protected] |
Wood 哥邮箱 |
| 授权码 | WUEw8qhBwjzpUAZW |
✅ 固化 |
| SSL 验证 | ssl=True |
必须 |
| 超时时间 | 60 秒 |
防超时 |
| ID 命令 | ASCII only! | ❌ 禁止中文 |
| 抓取方法 | BODY.PEEK[HEADER.FIELDS] |
✅ 已验证 |
| 删除方法 | set_flags + expunge |
✅ 已验证 |
🎯 关键词列表（固化）
可删除关键词（DELETE_KW）
DELETE_KW = [
'sale', 'discount', 'promo', 'deal', 'offer', 'clearance',
'newsletter', 'subscription', 'weekly', 'monthly', 'unsubscribe',
'notification', 'alert', 'update', 'reminder', 'verification',
'ecobee', 'rachio', 'nest', 'ring', 'smart home',
'hoa', 'community', 'meeting', 'election', 'board',
'temu', 'shein', 'wish', 'aliexpress', 'sponsor', 'ad'
]
保留关键词（KEEP_KW）
KEEP_KW = [
'forsyth', 'school', 'lhs', 'teacher', 'student',
'chase', 'visa', 'statement', 'bank', 'credit card',
'amazon', 'order', 'shipping', 'tracking', 'delivery',
'uber', 'lyft', 'flight', 'hotel', 'airline', 'delta',
'google', 'icloud', 'dropbox', 'onedrive', 'apple',
'insurance', 'medical', 'health', 'doctor', 'hospital',
'tax', 'irs', 'government', 'utility', 'power', 'water',
'receipt', 'invoice', 'warranty', 'contract', 'lease'
]
📋 标准使用流程（6 步法）
第 1 步：分析模式
python3 /Users/lobster/.openclaw/workspace/skills/email-analyzer/analyze.py \
--start-date "2021-02-26" \
--end-date "2021-08-26" \
--mode analyze
输出：JSON 分析报告（分类统计 + UID 列表）
第 2 步：等待用户确认
必须等 Wood 哥回复"删除"或"确认"才能继续！
汇报格式：
📊 邮件分析报告
批次：8
日期范围：2021-02-26 ~ 2021-08-26
总邮件数：1,500 封
建议删除：700 封（46.7%）
- Temu 推广：280 封
- 智能家居通知：150 封
- HOA 社区邮件：120 封
- 其他营销：150 封
建议保留：800 封
- 学校/教育：200 封
- 财务/银行：180 封
- 购物订单：220 封
- 其他重要：200 封
确认后回复"删除"执行删除操作。
第 3 步：备份
python3 /Users/lobster/.openclaw/workspace/skills/email-analyzer/backup.py \
--batch 8 \
--uids-file delete_uids.json \
--output batch8_delete_final_backup.json
第 4 步：执行删除
python3 /Users/lobster/.openclaw/workspace/skills/email-analyzer/delete.py \
--uids-file delete_uids.json \
--confirm
第 5 步：验证
python3 /Users/lobster/.openclaw/workspace/skills/email-analyzer/verify.py
输出：删除前后对比统计
第 6 步：汇报
✅ 批次 8 删除完成！
📊 删除统计：
批次 8: 700 封 (2021-02-26 ~ 2021-08-26)
总计：700 封
📈 删除后邮箱状态：
总邮件数：26,298 封
未读数：18,500 封
💾 备份文件：
- batch8_delete_final_backup.json
✅ Wood 哥，清理完成！
🚨 错误处理
IMAP 连接失败
错误："Unsafe Login" 或 "Authentication failed"
处理：
1. 等待 5 分钟重试
2. 检查 126 网页版是否能登录
3. 汇报 Wood 哥：可能需要重新生成授权码
SELECT 失败
错误："SELECT failed" 或 "Mailbox not found"
处理：
1. 等待 24 小时（可能是 126 风控）
2. 用网页版登录解锁
3. 重试
ID 命令错误
错误：UnicodeEncodeError
处理：确保 ID 命令参数只用 ASCII 英文
⚠️ 使用限制
- 禁止自动执行：只有在 Wood 哥明确要求时才运行
- 禁止修改配置：授权码、关键词、服务器等都锁死
- 禁止跳过确认：删除前必须等 Wood 哥确认
- 禁止删除备份：备份文件永久保留
📁 文件结构
/Users/lobster/.openclaw/workspace/skills/email-analyzer/
├── SKILL.md # 本文档（固化配置）
├── analyze.py # 分析脚本
├── delete.py # 删除脚本
├── backup.py # 备份脚本
├── verify.py # 验证脚本
└── email_analyzer.py # 核心模块（IMAP 连接）
📝 变更记录
| 日期 | 变更 | 批准人 |
|---|---|---|
| 2026-03-06 | 技能创建 | Wood 哥 |
固化完成！🔒
安全使用建议
This skill will connect to imap.126.com as [email protected] using a plaintext password embedded in the code and can permanently expunge emails. Don't install or run this unless you are explicitly the owner of that account and you trust the source. If you are the legitimate owner and want similar functionality: (1) remove the hard-coded credential and supply credentials via a secure secret store or environment variable, (2) run first in analyze-only mode and inspect the generated report before any deletion, (3) verify backups are created and stored securely before running deletions, (4) avoid enabling autonomous agent invocation or require explicit human confirmation each run, and (5) consider rotating the mailbox password after testing. If you are not the mailbox owner, do not install — the package includes direct access credentials for a third party which is inappropriate and risky.
功能分析
Type: OpenClaw Skill
Name: email-analyzer
Version: 1.0.0
The skill bundle contains hardcoded plaintext credentials, specifically an IMAP authorization code ('WUEw8qhBwjzpUAZW') for the email account '[email protected]', found in both email_analyzer.py and SKILL.md. While the scripts' logic for analyzing and deleting emails is consistent with the stated purpose and includes safety measures like mandatory user confirmation and backups, the exposure of these credentials poses a significant security risk as it allows unauthorized access to the specified mailbox.
能力评估
Purpose & Capability
The name/description (analyze and delete emails in [email protected]) align with the code: IMAP client usage, searching, backup and expunge operations. However, instead of requesting credentials from the environment, the skill embeds a plaintext password and a hard-coded email/server pair for a named individual ('Wood 哥'), which is unusual and sensitive even if the declared purpose is to manage that specific mailbox.
Instruction Scope
SKILL.md and the scripts limit operations to analysis, backup, confirmation and deletion and instruct waiting for a manual confirmation. But the code supports automated deletion via CLI flags (--confirm) and the agent platform allows autonomous invocation by default. SKILL.md uses hard-coded local paths (/Users/lobster/...), locked configuration text, and asserts this is the '唯一合法方式' — these broaden scope and reduce flexibility to safely adapt credentials or destinations. There are no instructions to exfiltrate data externally, but the delete capability combined with embedded credentials is high-impact.
Install Mechanism
There is no install spec (instruction-only install) and no external downloads. All code is included in the bundle, so nothing is pulled from remote URLs at install time.
Credentials
The skill requests no environment variables but contains a plaintext IMAP password in CONFIG (password: 'WUEw8qhBwjzpUAZW') and a hard-coded email account. Embedding credentials in the package is disproportionate and dangerous: it exposes sensitive access in the skill itself and prevents the user or platform from controlling which credentials are used.
Persistence & Privilege
always:false (good) but disable-model-invocation:false (default) means an agent could call this skill autonomously. Combined with embedded credentials and scripts that can perform irreversible deletion (expunge), the blast radius is significant if the skill is invoked without explicit human oversight. SKILL.md's 'locked' claims are not enforceable by the platform and shouldn't be relied on.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install email-analyzer - 安装完成后，直接呼叫该 Skill 的名称或使用
/email-analyzer触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.0
- Initial release of the Email Analyzer skill (固化版).
- Provides a secure, locked-down process for analyzing and cleaning the [email protected] email account.
- All configurations (server, account, authorization code, keywords) are fixed and modification requires approval.
- Includes step-by-step workflow: analyze, confirm, backup, delete, verify, and report.
- Error handling and strict usage limits are documented to ensure safe and authorized operation.
元数据
常见问题
Email Analyzer 是什么？
分析并清理 [email protected] 邮箱中的邮件，按固化关键词分类后生成报告，待确认后执行安全删除操作。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 224 次。
如何安装 Email Analyzer？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install email-analyzer」即可一键安装，无需额外配置。
Email Analyzer 是免费的吗？
是的，Email Analyzer 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Email Analyzer 支持哪些平台？
Email Analyzer 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Email Analyzer？
由 farmerboybi-bot（@farmerboybi-bot）开发并维护，当前版本 v1.0.0。
推荐 Skills