FraudTagger Apple Mail Integration

This repository contains all you need to allows Apple Mail to automatically check incoming emails using the FraudTagger API and flag potential spam/machine generated email addresses in the sender field.

Components

• fraud_check.sh: A shell script that queries the FraudTagger API.
• fraudtagger.scpt: An AppleScript that integrates the shell script with Apple Mail.

Setup Instructions

1. Copy fraud_check.sh to ~/bin/ and make it executable:

   chmod +x ~/bin/fraud_check.sh

2. Edit the script and insert your API key.

3. Copy fraudtagger.scpt to ~/Library/Application Scripts/com.apple.mail/

4. Edit the Applescript and insert your username in the line - do shell script "/Users/YOURUSERNAME/bin/fraud_check.sh " & quoted form of senderEmail.

5. In Apple Mail:

   • Create a rule that runs the script for every message.
   • The rule will flag emails if the classification is "SPAM".

API Key Usage

Paid Access (With API Key)

The script by default uses a paid API key for access. You can set your own API key in the script by modifying the API_KEY variable:

API_KEY="YOUR_API_KEY_HERE"

Free Tier Access (Without API Key)

You can test FraudTagger for free by commenting out the API_KEY variable:

#API_KEY="YOUR_API_KEY_HERE"

The API will allow a small number of checks per day (see latest website information or API response limits data for exact details)

API Access Terms

This repository provides integration code for the TQNN FraudTagger API. The client-side code is open-source and available under the MIT License. However, access to the API service is governed by our Terms of Service.

• Free Tier: You may make up to 10 API calls per day without a paid subscription.
• Paid Tiers: Additional API calls are available through subscription plans. See our pricing page for more details.
• API Key: You must register for an API key to use the service.

Notes

• The script logs API responses to ~/fraudtagger_log.txt.
• Customize the AppleScript to move flagged messages to "Junk" if desired.

📄 Example API Response

This is a typical JSON response returned by the fraud_check.sh script and logged to ~/fraudtagger_log.txt:

{
  "metadata": {
    "name": "TQNN Fraud Tagger",
    "endpoint": "/v1/scoreUsername",
    "method": "GET",
    "randomness_range": "0-100",
    "entropy_range": "0-4",
    "confidence_range": "0-100",
    "request_id": "6809c744af0f2",
    "timestamp": 1745471300,
    "runtime_seconds": 3.5673,
    "verbose_mode": "yes",
    "called_by": "Unknown AI system",
    "energy": {
      "energy_usage_kWh": "0.0001684572",
      "carbon_emissions_mg": "39.2505169829",
      "equivalent_meters_driven": "0.3270876415"
    }
  },
  "data": {
    "email": "charlotteedwards451@gmail.com",
    "score": 6,
    "score_engine": "V1.0.4",
    "release": "Beta",
    "randomness": "85.27",
    "entropy": "3.41",
    "confidence": "72.63",
    "classification": "SPAM",
    "recommended_action": "quarantine or flag for further investigation",
    "reasoning": {
      "spam_lists_check": "NOT_EVALUATED",
      "similarity_to_known_spam": "Some similarity to known spam",
      "randomness_score": "The randomness is quite high at (85.27) which is unusual for real names and words",
      "entropy_score": "High Entropy (3.41) (Mostly Unpredictable)"
    }
  }
}