Methods & Ethics
Stillford Research studies internet infrastructure and its abuse using data it collects itself. This page states what is collected, the limits the lab works within, and how data is handled and released. It follows the principles of the Menlo Report: respect for persons, beneficence, justice, and respect for law and public interest.
Data collection
The lab works from three kinds of data:
- Decoy services. Internet-facing systems that imitate real services and hold planted credentials that grant access to nothing. They record unsolicited connections and what those connections attempt.
- Public internet records. Zone files obtained under the terms of ICANN’s Centralized Zone Data Service, and the public Certificate Transparency logs.
- Openly listed files. Some servers used in attacks expose their files through directory listings that require no login. The lab finds these through third-party internet indexes and retrieves what is listed, to study the tooling attackers use.
No data from any employer, client, or private network is used.
Deception and harm minimization
Decoys exist only to observe access that nobody invited. They are built so that they cannot be used to reach or attack anyone else. The lab never uses credentials it captures, and never uses its planted credentials against third-party services.
The lab does not scan the internet itself, and it does not bypass authentication, guess credentials, or exploit vulnerabilities to reach data. It retrieves only what a server offers to any visitor. Retrieved files are treated as hostile: they are never executed outside isolated analysis, credentials found in them are never used, and data belonging to victims is not published.
Data handling and retention
Raw data stays on systems the lab controls. Material captured from attackers is treated as untrusted and is parsed only in isolated tooling.
Raw data is kept only as long as it is needed for ongoing research and for checking published results. Holdings are reviewed at least once a year, and data that is no longer needed is deleted or reduced to aggregates.
Anonymization of public releases
Public datasets use prefix-preserving anonymization for IP addresses. Zone data is published only as aggregate counts, never as lists of names. Every published figure carries its snapshot date and source, and what was observed is kept apart from what is inferred.
Notification
When the lab’s data indicates that a third party has been compromised, it decides case by case whether notification would help. When it would, the lab reports through the organization’s published security contact or the relevant national CERT.
Ethics review
Stillford Research is an independent research lab and has no institutional review board. It publishes these principles so that its practices can be examined, and it welcomes questions and challenges at security@stillfordresearch.org.
Data access requests
Raw data is available to vetted researchers on request, under a short data-use agreement. Send requests to contact@stillfordresearch.org with your affiliation and intended use.
Requests are reviewed by the lab’s founder. Expect a reply within two weeks.