DigiCert's Breach Waited 16 Months in an Internal System That a Server Migration Left Facing the Internet
DigiCert's Breach Waited 16 Months in an Internal System That a Server Migration Left Facing the Internet
DigiCert, one of the world's largest certificate authorities, is notifying customers of a data breach that began with a February 2025 server migration error — and sat undetected in the gap between "unauthorized accounts" and "accessed data" for four months. The company says the breach affected a small portion of its archived customer data: names, business email addresses, phone numbers, and business addresses — and for some customers, personal identification documents such as driver's licenses, passports, and utility bills submitted during certificate validation. DigiCert is still reviewing the identification documents to determine exactly who was affected, and it has not said how many people are impacted.
The anatomy comes from an incident report DigiCert filed on Mozilla's Bugzilla platform (DigiCert participates in Mozilla's root program). A February 2025 migration mistakenly exposed an internal business system to the internet that was previously reachable only over VPN. That system included a self-registration feature built for other applications — one that did not require MFA — allowing external users to create accounts. Those accounts were not assigned to any application, but an authorization flaw let them read archived customer information anyway, in read-only mode. The unauthorized account at the center of the case accessed that data between June 27 and July 2, 2026.
The detection failure is the story's sharpest edge. DigiCert first noticed unauthorized accounts on the exposed system in June 2026, but — in its own words — "the incident was initially handled as an account-creation issue rather than a data-access event." The account was blocked on July 2. It took a later report through the company's managed security research program, renewed suspicion on October 12, and a full investigation before October 19 confirmed that customer data had actually been accessed. Since mid-October the company has restricted access to the affected system, disabled the self-registration feature, verified or removed remaining non-DigiCert accounts, reset passwords for all business system users, and added monitoring and alerting — remediations that, done at migration time, would have cost far less.
The blast radius has hard limits: the account had read-only access and could not modify records, approve certificate validations, or issue certificates. DigiCert says its certificate authority services, CertCentral, and DigiCert ONE platforms were not compromised, and that customer login credentials and issued certificates remain secure. Identity documents in the archived data, however, are exactly the material that fuels targeted phishing against the businesses that trusted the CA with them — one affected customer, cloud architect Fabian Bader, was blunt about it: "Thanks @digicert that your data leak was 'limited' to only my full PII data and potentially any verification documents I submitted for my certificate."
What Should You Do?
- If you receive a DigiCert notification, take the ID-document exposure seriously. Passports, licenses, and utility bills attached to your business identity are phishing and impersonation fuel — brief the people whose documents were involved.
- Verify your certificate estate anyway. DigiCert says issuance and validations were untouched, but this is a free moment to confirm your certificates, DNS validation records, and CertCentral account settings have not drifted.
- Harden the account. Turn on MFA for your DigiCert account, rotate its password, and review authorized users — the same housekeeping the breach's self-registration hole exploited.
- Run your own migration audit. Inventory what changed exposure during your last infrastructure migration: systems that moved from VPN-only to internet-facing, features that were "temporary," and authz models that were never designed for untrusted networks.
The WAF Angle
Strip away the CA brand name and this is OWASP's number one category — broken access control — wearing its most common disguise: an internal system that infrastructure change quietly turned into an internet-facing app, complete with a self-registration feature and an authorization model nobody re-tested for the new exposure. The lesson for web defenders is that "internal" is a network fact, not a property of the application, and every migration is a re-deployment that deserves the same authz review as a new release. The second lesson is detection taxonomy: DigiCert saw unauthorized accounts in June and classified them as an account-creation nuisance; the data access underneath went unnoticed for months. Anomalous account creation on a system that shouldn't accept outside registrations is a high-severity signal by default, and classification discipline — treating every identity anomaly as a potential access-control event until proven otherwise — is what separates the teams that catch these from the ones that file them away. The same broken-access-control pattern drove Drupal's actively exploited Commerce Core bypass and Quest's unauthorized-access disclosure this quarter.