email-intelligence（电子邮件情报）

Email Intelligence

Analyze your email inbox health using creative weather metaphors, intelligent classification, debt scoring, and ghost detection. Transform overwhelming inbox analysis into clear, actionable insights.

When to Use

Use this skill when the user asks about:

• Inbox status or health
• Email overwhelm or management
• Who they're "ghosting" (ignoring responses to)
• Time cost of processing their inbox
• Signal vs noise in their email
• Email debt or backlog

Requirements

• himalaya CLI configured with IMAP access
• Python 3.6+
• The configured email account should have standard folder names (INBOX, Archive)

Quick Usage

# Basic analysis (last 7 days)
python3 scripts/email_classify.py

# Extended analysis (last 14 days)
python3 scripts/email_classify.py --days 14

# JSON output for integration
python3 scripts/email_classify.py --format json

Weather System

The inbox "weather" is determined by the number of human emails in your INBOX that need responses:

• 🌊 Calm Seas (0-2): Inbox is peaceful and manageable
• 🍃 Light Breeze (3-5): A few emails need attention, nothing urgent
• 🌬️ Choppy Waters (6-10): Multiple emails require responses
• ⛵ Small Craft Advisory (11-20): Many people waiting for replies
• ⛈️ Storm Warning (21+): Inbox is overwhelming, needs immediate attention

Email Classification

Emails are automatically classified into four categories:

🤖 Automated

• Sender contains: noreply, no-reply, donotreply, notifications@, alerts@
• Subject contains: unsubscribe, automatic, auto-generated
• Time cost: 0 minutes (can be safely ignored)

📰 Newsletter

• From domains: substack.com, email.*, marketing.*, updates@
• Subject contains: newsletter, digest, weekly, monthly, roundup
• Time cost: 1 minute each (skim or archive)

🔔 Notification

• From services: github, slack, jira, linear, aws, google, microsoft, etc.
• Time cost: 30 seconds each (quick action/acknowledgment)

👥 Human

• Everything else (actual people writing)
• Time cost: 3 minutes each (requires thoughtful response)

Metrics Explained

Email Debt Score (0-100)

Calculated based on unseen human emails in your INBOX, weighted by age:

• < 1 day old: 1 point each
• 1-3 days old: 3 points each
• 3-7 days old: 5 points each
• 7+ days old: 10 points each

Score meanings:

• 0-30: 🟢 Great job! Inbox under control
• 31-60: 🟡 Getting busy. Consider tackling some replies
• 61-100: 🔴 High debt! Many emails waiting for responses

Signal-to-Noise Ratio

Percentage of emails that are from humans (signal) vs automated/newsletter/notification (noise).

Higher ratio = more meaningful email, less spam/clutter.

Ghost Report

Shows up to 5 people you're "ghosting" (human emails you haven't read), sorted by how long they've been waiting. Helps prioritize who needs responses most urgently.

Time Investment

Estimates how much time you'll need to process your current inbox based on email types and their typical processing times.

Output Formats

Text Format (Default)

Human-readable report with weather, debt score, signal/noise analysis, ghost report, and time estimates. Perfect for quick status updates or daily reviews.

JSON Format

Structured data for integration with other tools, APIs, or dashboards:

{
  "weather": {
    "level": "breeze",
    "emoji": "🍃", 
    "description": "Light Breeze - A few emails need attention...",
    "humanCount": 4
  },
  "categories": {
    "automated": 15,
    "newsletter": 8,
    "notification": 12,
    "human": 4
  },
  "debtScore": 18,
  "ghostReport": [...],
  "signalNoiseRatio": {
    "ratio": 0.103,
    "percentage": "10%"
  },
  "timeCost": {
    "minutes": 18,
    "formatted": "18 minutes"
  }
}

Integration Tips

Dashboard Integration

The JSON output is designed for easy integration into dashboards or status displays. The weather metaphor makes it intuitive for users to understand their inbox state at a glance.

Automation

Run periodically (via cron) to track inbox health trends over time. The debt score is particularly useful for identifying when inbox maintenance is needed.

Alerts

Set up alerts when:

• Weather reaches "Storm Warning" level
• Debt score exceeds 60
• Ghost report shows people waiting >5 days

Troubleshooting

No emails found

• Verify himalaya CLI is configured: himalaya envelope list -f INBOX --page-size 1
• Check folder names match your email provider (some use "Inbox" vs "INBOX")

Classification seems wrong

The classification uses heuristics based on sender patterns and subject lines. You can modify the patterns in the script for your specific needs.

Slow performance

• Reduce the --days parameter to analyze fewer emails
• The script fetches up to 200 emails from Archive - adjust page_size in the code if needed

Philosophy

Email Intelligence treats your inbox as a living system with its own weather patterns and health metrics. Rather than just counting unread emails, it helps you understand:

• What actually needs your attention (humans vs bots)
• How "behind" you are (debt score)
• Who you might be accidentally ignoring (ghost report)
• How much time you're really looking at (time cost)

The weather metaphor makes inbox status intuitive and actionable - you know to "batten down the hatches" when a storm is brewing, but you can relax when the seas are calm.

Use this skill to transform inbox anxiety into clear, prioritized action plans.