URL Classification
Discovered Intelligence - Dissemination
- Phishing is a threat that impacts the entire financial/banking industry, so any research done and/or tools created by a company should be shared amongst all companies to help work towards minimizing these threats.
- As a phishing attack on any employee can impact a whole organization, all employees should be made aware of phishing threats and how to detect them.
- Through the findings, guides can be created that are shared amongst all employees of companies outlining common trends/patterns to look for in links before clicking.
- With malicious URLs being everchanging, employees should report any suspicions from previously clicked links and/or anytime they receive a potentially suspicious link.
Discovered Intelligence - Courses of Action
- To help protect organizations as a whole, companies can:
- Implement 3rd-party (and/or develop in-house) tools/software that scan links prior to being opened and/or loading the page.
- Create stricter criteria for email filters.
- To help ensure employees remain vigilant to phishing threats, companies can implement internally planned/safe phishing exercises.
- Those who report or don't click the links are rewarded whereas those who do click are assigned additional training.
- This helps companies monitor awareness of these threats among their employees and adjust trainings if/where necessary.
Discovered Intelligence - Iterative Changes
- With the dataset for non-malicious URLs being limited to 500 URLs, we would like to find a source that provides more data to create a larger dataset and even out the balance between malicious vs. non-malicious URLs.
- While the datasets provided are more generalized URLs (vs. industry specific), we would like to find datasets that are specific to the financial industry.
- This will help provide more detail to the analysis and potentially uncover additional characteristics/themes in the URLs.
- Ensure the data from both sources is similar and not missing certain characteristics.
- It was noticed that not all the URLs provided by Moz.com had HTTP/HTTPS and www present which creates potential discrepancies.
- Additionally, URLs provided by Moz.com did not contain URL parameters.