The ANTS confirms a security incident: a DMARC configuration inactive since at least 2019

The ANTS portal (ants.gouv.fr) experienced a confirmed security breach on April 15, 2026, as reported by the Ministry of the Interior. Data was exfiltrated, including names, email addresses, and dates of birth. Beyond the technical flaw, this incident reveals two significant vulnerabilities: a failing dark web alert qualification process and an email configuration that has been neglected since at least 2019.

The National Agency for Secure Titles, established in 2007 and now rebranded as France Titres, is the public agency responsible for overseeing all official French documents, including national identity cards, biometric passports, driving licenses, and vehicle registration certificates. Anyone who has renewed one of these documents in France over the past decade has a personal account on the ANTS portal, linked to their name, email address, date, and place of birth. This database of sensitive information was compromised in the April 15 incident. On April 20, the Ministry of the Interior confirmed that a security breach had been detected, potentially affecting data from both private and professional accounts. However, attachments submitted during the application processes, including identity documents and proof of residence, were reported to be unaffected, and the leaked data would not provide unauthorized access to accounts.

Regulatory measures were initiated immediately. A notification was submitted to the French Data Protection Authority (CNIL) under Article 33 of the GDPR, a report was sent to the Paris public prosecutor in compliance with Article 40 of the Code of Criminal Procedure, and an alert was issued to the National Cybersecurity Agency (ANSSI). The Anti-Cybercrime Office is now involved. Affected users are being provided with personalized notifications. A malicious actor using the alias “breach3d” claims on cybercrime forums to be selling 19 million records, although this figure has not yet been confirmed by authorities as investigations continue into the extent of the breach.

Prior Warning in September 2025

The security timeline for the ANTS portal does not begin on April 15, 2026. In September 2025, the agency was already mentioned in relation to an alleged civil registry data leak. A database containing 12 to 13 million records was reportedly put up for sale on the dark web, purportedly sourced from the agency. However, investigations at that time, led by ANSSI, found no evidence of an intrusion into the agency’s systems, and the data in question exhibited numerous inconsistencies with ANTS’ internal formats. This position was reaffirmed in an official update on March 6, 2026.

On April 15, another incident was detected on the portal. Whether the two events are linked to a single late-detected compromise or represent two separate intrusions remains uncertain as investigations are ongoing. What this sequence illustrates for Chief Information Security Officers (CISOs) is that a documented dark web alert from September 2025 did not prompt an intrusion detection, which was confirmed only seven months later. Security researchers identified the attack vector as an Insecure Direct Object Reference (IDOR) vulnerability in the portal’s API, allowing access to other users’ profiles by simply modifying a numeric identifier in the requests. However, this incident has neither been confirmed nor denied as investigations are still in progress.

DMARC in Monitoring Mode Since at Least 2019

In this context, the concern raised by Loïc Guezo, the cybersecurity strategy director at Proofpoint and vice president of CLUSIF, takes on significant meaning. The DMARC (Domain-based Message Authentication, Reporting, and Conformance) record for the domain ants.gouv.fr is set to p=none, which puts it in monitoring mode—collecting reports without enforcing any policies. Consequently, a fraudulent email appearing to come from [email protected] generates normally in recipients’ inboxes with neither quarantine nor rejection. This configuration has been documented since at least July 2019. The sp=none setting extends the same lack of protection to all subdomains.

The compromised civil data represent a valuable target, as they can facilitate targeted phishing campaigns. A personalized email using the recipient’s name, sent from an unfiltered ants.gouv.fr domain, requesting action on a driving license or vehicle registration case, presents a level of credibility that no spam filter can detect when domain authentication is absent. Businesses included in the data, such as driving schools, car dealerships, and accredited service providers, face an additional risk: their SIREN numbers and credential identifiers can be exploited for highly effective commercial identity theft.

The Challenge of Qualifying Dark Web Alerts

ANTS qualifies as an essential entity under the NIS2 Directive, which has been transposed into French law and imposes specific obligations for detecting and responding to incidents. Notifications to the CNIL and reports to the public prosecutor were made in compliance with the GDPR. However, these procedural obligations do not address the critical issue highlighted by this sequence: the ability to accurately qualify a dark web alert before the reality of an intrusion is technically confirmed.

For CISOs managing organizations of comparable criticality, two operational avenues stand out from this incident. First, regular reviews of email authentication configurations, particularly DMARC, are a foundational measure, becoming indefensible when large-scale civil data is at stake. Second, establishing an internal protocol for responding to dark web alerts—separate from the technical confirmation process for an intrusion—conditions the ability to minimize the exposure window between the initial signal and an operational response. This interval is precisely what the ANTS sequence exemplifies over seven months.

Scroll to Top