document-extractor

name	document-extractor
description	Extracts key information from documents (PDFs, images, text files) in a folder and creates structured markdown summaries with the same filename but .md extension. Use when you need to analyze and extract insights from multiple documents in a directory.
allowed-tools	Read, Write, Bash, Glob
category	documents-and-data
metadata	{"version":"1.0","changelog":["1.0: Promoted to trinity-skills library (2026-08-04)"]}

Document Extractor

Analyze documents in a specified folder and create structured markdown extracts that capture the most important information from each document.

Workflow

1. Initial Setup

• Get current date/time: date '+%Y-%m-%d %H:%M:%S %Z'
• List all files in the specified folder
• Create output directory: Files/[folder_name]/ in the workspace

2. Document Processing

For each document:

1. Read using the Read tool
2. Collect metadata (file hash, modification date, extraction timestamp)
3. Identify document type and extract key information
4. Write to Files/[folder_name]/[original_name].md

3. Information Categories

Legal/Official Documents: Document type/number, dates, issuing authority, key terms, obligations, deadlines

Financial Documents: Account numbers (masked), transactions, amounts, payment dates, tax info

Insurance Documents: Policy numbers, coverage types/limits, premiums, deductibles, exclusions

Contracts/Agreements: Party names, duration, obligations, payment terms, termination clauses

Personal Documents: Document numbers (masked), validity periods, restrictions

4. Output Template

# [Document Title/Type]

## Document Information
- **Document Type**: [Type]
- **File Name**: [Original filename]
- **File Hash**: [MD5/SHA256 hash]
- **Last Modified**: [File modification date]
- **Extraction Date**: [Current timestamp]

## Key Information
[Main extracted data organized by relevance]

## Important Dates
- **Issue Date**: [Date]
- **Expiry Date**: [Date if applicable]
- **Next Action Required**: [Date/Description if applicable]

## Notes and Observations
- [Important observations]
- [Unusual clauses]
- [Action items]

---
*Extracted: [timestamp] | Original: [path]*

5. Security

• Mask sensitive info (show last 4 digits only)
• Note when information has been redacted
• Flag potential security concerns

6. Summary Report

After processing all documents, create Files/[folder_name]/summary.md with:

• Total documents processed
• Document categories found
• Key findings
• Items requiring attention

Useful Commands

# Get timestamp
date '+%Y-%m-%d %H:%M:%S %Z'

# Get file hash (macOS)
md5 -q /path/to/file
shasum -a 256 /path/to/file | cut -d' ' -f1

# Get modification date (macOS)
stat -f "%Sm" -t "%Y-%m-%d %H:%M:%S" /path/to/file

# List documents
find /path/to/folder -type f \( -name "*.pdf" -o -name "*.txt" -o -name "*.jpg" -o -name "*.png" \)

Error Handling

Error	Cause	Solution
File unreadable	Corrupted or encrypted file	Skip file, note in summary, try alternative viewer
PDF extraction fails	Scanned image PDF	Use OCR or note as "image-based PDF"
Output folder not writable	Permission issue	Check permissions on Files/ directory
Sensitive data detected	PII or credentials visible	Mask appropriately, flag for review
Duplicate files	Same content, different names	Note duplicates in summary, extract once
Unknown file type	Unsupported format	Skip with note, or convert to supported format first