phishradar 0.1.post2

Phishing domain detection from Certificate Transparency logs.

PhishRadar

Detect phishing/scam domains from Certificate Transparency logs. Can be useful for internal security teams for monitoring company-specific domains, as well as for sectorial/national CERTs for monitoring the domains of their constituents.

Logs are retrieved via Certstream. Keywords are set in a YAML configuration file and detected via two methods:

• for common subwords, results from WordSegment are scanned
• for rarer words (e.g. brand names), we simply check if the domain contains the keyword - this is because WordSegment may not correctly separate unknown subwords.

For better customization, a minimal threshold of matching keywords can be set.

The idea was born after encountering phishing attacks and malware Command-and-Control communication involving domains impersonating Armenian government bodies (1, 2).

Installation and usage

pip install phishradar
phishradar --config ./config.yaml

Sample configuration

certstream_url: wss://certstream.calidog.io/
keywords:
  - test
  - keyword
  - yourbrand
whitelist:
  - exclude
  - these
threshold: 1
output:
  console: True
  file: output.log

Further work

• Implement webhook sinks
• Try out alternative threshold mechanism (e.g. weighted keywords)
• Experiment with word segmentation via a Small Language Model