Definition
Collecting host certificates from network traffic or other passive sources like a certificate transparency log and analyzing them for unauthorized activity.
How it works
Certificates are analyzed outside of a TLS server connection using third-party secure update logs, domain name analysis and analytics.
Secure update certificate logs
The key enabling feature is a secure service that maintains record logs of certificate activities. The logs allow users to only append certificates and never to delete or modify the log entries. The logs use Merkle Tree Hashes to ensure they have not been tampered with. The logging service also allows for public auditing by any user.
The logging service, upon receipt of a certificate to log, will respond with a signed certificate timestamp (SCT). The SCT guarantees the certificate will be added to the log within the time specified. The SCT must be present with the certificate during a TLS handshake.
Certificate monitoring, of the logs, is typically done by the CA and they watch for suspicious certificate logging and unusual certificates or extensions or permissions. Monitors are also responsible for verifying the logs are accurate and public.
Log integrity is verified by log auditors. Auditors make use of log proofs are used to validate the cryptographic hashes (Merkle Trees) that the log employs are consistent. In order to ensure consistency throughout multiple monitors and auditors, sharing a common logging service, gossip protocol is employed.
- Certificate Logs
- Certificate Monitoring
- Certificate Auditors
Phishing domain name analysis
- A curated corpus of known benign domains and phishing domain names is used as training text for machine learning. Through the use of feature set extraction, vectors labels are created with scoring to indicated if they are considered benign or phishing domains.
- A stream of new or updated SSL certificates with fully qualified domain names (FQDN) is analyzed against the feature vectors and a predictive model determines a score for the domains. The scoring considers distance measures such as Levenshtein distance to help in determining the final label score. Supervised learning is also employed using the curated domains of benign and phishing domains.
- Subdomain phishing analysis, prepending a trusted domain to a phishing domain, and regular expression comparisons are also used in the label scoring model. A tunable measure is used to determine the threshold for alerting. This measure helps to balance between precision and recall measures.
Considerations
- Some entity will need to run the logging service and a trusted entity is preferred.
- Certificate Authorities will likely need to monitor the logging service for consistency.
- Certificate revocation is unchanged and remains outside of Certificate Transparency, but certificates needing to be revoked are visible.
- Technique dependent of reliable feed of new and updated certificates
- Some certificate authorities allow for certificates to be registered with wildcards in the FQDN and thus will fail some of the subdomain scoring
- Phishing HTTP domains will not be discovered
Implementation perspective
Passive Certificate Analysis should be treated as a technical defensive capability rather than a product checkbox. In practice, teams should define the protected scope, the conditions under which the technique acts, and the observable evidence that demonstrates the intended behavior. For this technique, likely engineering context includes Certificate File.
Use the technique to identify suspicious, unauthorized, or abnormal activity through observable evidence and repeatable analysis.
Questions to ask
- Which events, states, or artifacts must be observed for the analysis to work?
- What analytic logic, threshold, comparison, or signature turns observations into a finding?
- How are expected false positives, blind spots, and environmental variations documented?
- Who receives the result, and what action is expected when the technique produces a finding?
Evidence and validation
- Telemetry and data-source configuration records
- Analytic logic, thresholds, signatures, and version history
- Test cases demonstrating expected positive and negative results
- Alert, triage, escalation, and tuning records
Common failure patterns
- Required telemetry is missing, delayed, or transformed in a way that invalidates the analysis.
- The technique produces alerts without an accountable triage and response process.
- Detection coverage is claimed from product deployment without testing the relevant analytic behavior.
This implementation perspective is original Bare Metal Cyber educational content. It does not replace the official D3FEND definition or establish that a specific product implements the technique.
Technique hierarchy
Top-level family
Parent techniques
Direct child techniques
None listed at this level.
Artifacts and ontology entities
These relationships describe how D3FEND connects a defensive technique to artifacts or other ontology entities. They describe graph semantics, not a product certification.
Show inferred artifact relationship paths (1)
Offensive-technique relationships
These relationships are generated from D3FEND graph paths and are explicitly experimental. They should be treated as hypotheses for defensive analysis—not as proof that the technique prevents, detects, or removes an offensive behavior.
ATT&CK Enterprise (6)
Authoritative sources
- Open this technique on the official D3FEND website ↗
- Open the official ontology resources ↗
- Certificate Transparency ↗Google · Google · TechniqueReference
- StreamingPhish ↗Uber · Wes Connell · TechniqueReference
Bare Metal Cyber is an independent educational publisher and is not affiliated with or endorsed by The MITRE Corporation. MITRE D3FEND™ and the D3FEND logo are trademarks of The MITRE Corporation. MITRE ATT&CK® and ATT&CK® are registered trademarks of The MITRE Corporation. Use of D3FEND source material is subject to the official Terms of Use.