The Digital Agency says personal information may have leaked after attackers exploited a VPN vulnerability. The agency is treating around 246,000 records as potentially affected while its investigation continues. It has not said that every record was definitely stolen.
Japan's Digital Agency disclosed on 11 September 2026 that its Government Solution Service, known as GSS, had been hit by unauthorised access linked to a vulnerability in a VPN device.
The agency says it first detected unusually large-scale file access on 25 June 2026. The activity was carried out using an account belonging to a maintenance and operations worker, prompting an investigation.
On 9 July, investigators concluded that a third party had exploited a vulnerability in a network connection device described by the agency as a VPN. The affected maintenance account was suspended and external communications with the compromised equipment were blocked.
The affected VPN equipment was part of the infrastructure used to provide external connectivity to Japan's Government Solution Service. The Digital Agency has not publicly named the VPN vendor, product or specific CVE.
The VPN flaw was already known before it was exploited
One of the most important details in the Digital Agency's disclosure is that the vulnerability was not an undisclosed zero-day. The agency says information about the flaw had already been made public before the attack.
According to the agency's Q&A, the vulnerability was initially assessed as Medium under CVSS. The Digital Agency says it had started responding to the issue more quickly than would normally be expected from the published severity rating, but the vulnerability was exploited before the corrective patch was applied.
What Japan has disclosed about the vulnerability
- The flaw affected a VPN network connection device used by GSS.
- The vulnerability had been publicly disclosed before the attack.
- Its original severity assessment was Medium under CVSS.
- The patch had not yet been applied when the vulnerability was exploited.
- The Digital Agency has not named the VPN vendor, product or CVE.
The agency says it is withholding more technical detail because revealing the exact weakness could interfere with ongoing security protections. It has also declined to publish detailed information about GSS's internal security design.
Accuracy note: the Digital Agency has not publicly identified the VPN manufacturer or vulnerability number. Claims naming a specific product or CVE should therefore be treated cautiously unless the agency later confirms them.
What information may have been exposed?
The Digital Agency says files potentially accessed by the attacker contained personal information relating to staff and other people connected with organisations using GSS. The total is approximately 246,000 records.
That figure is split into around 189,000 records relating to staff of GSS-using organisations and public-sector personnel involved in their work, plus around 57,000 records relating to businesses and individuals involved in work for those organisations.
Names
Around 236,000 name records are within the potentially affected files.
Email addresses
Around 231,000 email address records may have been affected.
Phone numbers
Around 94,000 phone number records were identified in the potentially affected data.
Addresses
Around 1,000 address records may have been affected. The agency says many contact details were work rather than private details.
The categories overlap, so those figures should not be added together as separate individuals. The Digital Agency also says many of the phone numbers and addresses were government workplace contact details rather than personal home information.
Data the agency says was not included
- Japan's My Number identification numbers.
- Bank account information.
- Pension numbers.
- Personal information belonging to the general public unrelated to GSS work.
As of the agency's disclosure, it said it had found no confirmed secondary misuse of the information linked to the incident. However, it warned that exposed contact details could be used in impersonation or phishing attempts.
Japan Digital Agency VPN breach timeline
The incident was detected in June, the VPN entry point was confirmed in July, and the potential personal-data exposure was publicly disclosed in September.
- Large-scale file access detected The Digital Agency detects access to a large number of server files using a maintenance and operations account and begins investigating.
- VPN vulnerability identified as the entry route Investigators determine that a third party exploited a vulnerability in VPN equipment. The relevant account is suspended and external communications with the compromised equipment are blocked.
- Privacy regulator notified The Digital Agency reports the incident to Japan's Personal Information Protection Commission after identifying a risk that personal information may have leaked.
- Potential exposure of around 246,000 records disclosed The agency publicly announces the incident, the categories of information potentially affected and the initial containment measures.
- Incident Q&A updated The Digital Agency adds further detail, including the regulator notification, the known status of the VPN vulnerability and the scale of the affected data.
What has Japan's Digital Agency done in response?
The Digital Agency says it applied the VPN patch, changed authentication information for related accounts and blocked external communications involving the compromised equipment.
It says no further unauthorised access or suspicious communications have been observed since those containment measures were put in place.
- Patched the VPN vulnerability The corrective update has now been applied to the affected equipment.
- Suspended and changed affected account credentials The maintenance account involved in the incident was stopped and related authentication information was changed.
- Blocked external communications Communications between the compromised equipment and external systems were cut as part of containment.
- Reviewing vulnerability management The agency says it will strengthen how vulnerabilities are assessed and handled, with more rapid action based on the real-world risk to government systems.
- Reviewing external access methods Japan says it is also improving how outside connections into the government environment are controlled.
The agency says the response did not disrupt government work carried out through GSS. It has also begun contacting affected people individually as their details are identified.
Why this VPN breach matters
The incident is notable because the weakness was already known and had only been assigned a Medium severity rating. In this case, a vulnerability that may not have appeared critical on its score alone still became an entry point into a major government environment before patching was completed.
| Question | What Japan has confirmed | What remains unknown |
|---|---|---|
| How did attackers get in? | A vulnerability in VPN equipment was exploited. | The exact product, vendor and CVE have not been disclosed. |
| Was the flaw already known? | Yes. It was public before the attack and originally rated Medium. | The exact publication date and vulnerability identifier are withheld. |
| Was data definitely stolen? | The agency found files that may have leaked and is treating them as potentially affected. | It has not said that every one of the roughly 246,000 records was definitely exfiltrated. |
| Has the problem been contained? | The patch was applied, credentials changed and external communication blocked. | The wider investigation is continuing and details may still change. |
Our view: severity scores should not replace context
The important lesson is not that every Medium-rated VPN flaw should automatically be treated as critical. It is that vulnerability scores are only one part of a risk decision. An internet-reachable device protecting access to a high-value government environment can justify faster action than its headline CVSS category might suggest.
The Digital Agency itself says it is reviewing vulnerability management so it can respond more quickly based on the practical risk to government information systems. That is the most significant security takeaway from the incident.
Does this mean VPNs are unsafe?
No. This incident concerns a vulnerability in a specific, undisclosed VPN appliance used as part of a government network. It does not show that VPN technology as a whole is unsafe, nor does it indicate that consumer VPN apps were involved.
It does show why VPN gateways and other internet-facing remote-access systems need rapid patching, restricted administrative access, monitoring and controls that limit what a compromised account can reach.
What should potentially affected people watch for?
Japan's Digital Agency specifically warns about phishing and impersonation. People contacted in relation to the breach should be cautious with unexpected emails, calls or SMS messages claiming to come from the Digital Agency or another government organisation.
The agency says it will not ask people by email or phone to provide passwords, authentication credentials or credit card details. Unexpected links and attachments should therefore be treated cautiously.
Frequently asked questions
What happened in the Japan Digital Agency VPN breach?
Japan's Digital Agency says a third party exploited a vulnerability in VPN equipment used by the Government Solution Service. Investigators detected large-scale file access carried out through a maintenance and operations account.
How many records may have been exposed?
Around 246,000 records containing personal information are being treated as potentially leaked. The agency uses this figure because it cannot rule out leakage of the affected information; it has not said every record was definitely exfiltrated.
What information may have been leaked?
Potentially affected information includes around 236,000 names, 231,000 email addresses, 94,000 phone numbers and 1,000 addresses. These categories overlap. The agency says My Number IDs, bank account information and pension numbers were not included.
Was the VPN vulnerability a zero-day?
No. The Digital Agency says the vulnerability had already been publicly disclosed before it was exploited and was initially rated Medium under CVSS. The patch had not yet been applied when the attack occurred.
Which VPN product or CVE was involved?
The Digital Agency has not named the VPN manufacturer, product or CVE. It says more detailed disclosure could create additional security risk.
Were ordinary members of the public affected?
The Digital Agency says personal information belonging to the general public was not included. The affected information relates to staff, public-sector personnel, contractors, businesses and other individuals involved in work for organisations using GSS.
Has the VPN vulnerability now been patched?
Yes. The agency says the corrective patch has been applied, related account credentials were changed and external communications with the compromised equipment were blocked. It says no further unauthorised access or suspicious communications have been detected since containment.
