AI Skill Report Card
Scanning Email Content
Quick Start
Extract structured information from raw email text:
From: Simon Willison
Subject: How StrongDM's AI team build serious software without even looking at the code
Date: 17:54
From: ByteByteGo
Subject: EP201: The Evolution of AI in Software Development
Content: AI has fundamentally changed how engineers code. This shift can be described in three waves.
Date: 17:46
Recommendation▾
Add more concrete input/output examples showing messy Gmail export text with all the HTML artifacts, unicode characters, and interface elements that need cleaning
Workflow
- Identify Email Boundaries - Look for sender patterns, timestamps, and subject separators
- Extract Key Fields - Sender, subject, preview text, timestamps
- Clean Content - Remove visual noise (͏ characters, unsubscribe links)
- Categorize by Type - Newsletters, alerts, promotions, personal messages
- Structure Output - Present as clean list or table
Progress:
- Parse sender information
- Extract subject lines
- Clean preview content
- Identify timestamps
- Remove UI artifacts
- Format structured output
Recommendation▾
Include a template or framework for the structured output format (JSON schema, table format, etc.) rather than just describing it
Examples
Example 1:
Input: unread, Simon Willison from., How StrongDM's AI team build serious software without even looking at the code, 17:54
Output:
- Sender: Simon Willison
- Subject: How StrongDM's AI team build serious software without even looking at the code
- Status: Unread
- Time: 17:54
Example 2:
Input: ByteByteGo, Top Authentication Techniques to Build Secure Applications, 5 Feb, In this article, we'll explore multiple authentication techniques
Output:
- Sender: ByteByteGo
- Subject: Top Authentication Techniques to Build Secure Applications
- Date: 5 Feb
- Preview: In this article, we'll explore multiple authentication techniques
Recommendation▾
Expand the workflow with specific techniques for handling edge cases like forwarded emails, email threads, and non-English content
Best Practices
- Preserve original sender names and subjects exactly
- Convert relative timestamps to absolute dates when possible
- Group related emails (same sender, thread)
- Flag security alerts and important notifications
- Maintain chronological order
Common Pitfalls
- Don't lose information during cleanup
- Don't assume consistent formatting across emails
- Don't merge separate emails into one entry
- Don't ignore metadata like read/unread status
- Don't remove legitimate special characters from subjects