This guide creates a review queue from FullHunt’s potential-phishing and typosquatting datasets. It stores previously observed records so each run reports only new candidates.

Requirements

  • Enterprise account with the Dark Web Monitoring module
  • Python 3.10 or later
  • requests installed
  • FULLHUNT_API_KEY and FULLHUNT_MONITORED_DOMAIN environment variables
Both endpoints filter with the query parameter q. The potential-phishing route returns up to 1,000 items per page; typosquatting returns 10.

Run the monitor

brand_monitor.py

Production behavior

The first run treats every returned record as new. To create a silent baseline, run once, review the state file, and begin alert delivery on the next scheduled run. Use the output as an analyst queue, not an automatic phishing verdict. Enrich new candidates with WHOIS, certificates, and passive DNS. Add a disposition such as new, investigating, benign, confirmed, or takedown_requested in your case system.

Sensitive-data controls

  • Keep API keys and state files out of source control.
  • Restrict access to candidate and credential-exposure data.
  • Do not print raw passwords or unrelated personal fields.
  • Set retention based on incident-response and privacy requirements.