✦ Preferences saved
TRACE / CPR / NATIONAL IDENTITY DATA BREACH / OPEN CASE
07 OCTOBER 2026 / OPEN EVIDENCE FILE

NAMES, ADDRESSES, CPR NUMBERS AND OTHER INFORMATION WERE STOLEN FROM ABOUT 8.8 MILLION PEOPLE.

THE BILL CAUGHT IT. THE SECURITY CONTROLS DID NOT.

Criminals used a private Danish company's authorized CPR access to steal names, addresses, CPR numbers and other identity data on about 8.8 million people, and the illegal access ran for around ten days. On 6 October the ministry confirmed that an abnormally large bill exposed the scale before the security controls did. The authorities refuse to identify the company and still have not published the product, credential path, request telemetry, alert history or a person-by-person map of whose data was stolen.

WHY I AM ON THIS I LIVE UNDER THIS SYSTEM TOO.

My CPR is part of the same national identity infrastructure. I earn my living outside IT, pay for my own tools and do this investigation independently. I do not have a ministry, a press office or a consultancy budget behind me. I have their documents, their timestamps, their rules and the evidence I can preserve.

8.8MPEOPLE WHOSE DATA WAS STOLEN
~10DAYS THE ILLEGAL ACCESS RAN
02 OCTCPR FINALLY NOTICED THE ACTIVITY
NOAUTHORITIES NAME THE COMPANY?
PRIVATE COMPANY'S CPR ACCESS ABUSED
14M+ AUTOMATED CPR LOOKUP ATTEMPTS REPORTED
VALID CPR NUMBERS IDENTIFIED
NAMES + ADDRESSES + CPR NUMBERS STOLEN
THE BILL EXPOSED THE ACTIVITY AFTER ~10 DAYS
01 / CONFIRMED RECORD

THE DATA WAS STOLEN, AND THE TRAIL THAT SHOULD EXPLAIN TEN DAYS OF FAILURE IS STILL NOT PUBLIC.

What is public now is blunt enough on its own: names, addresses and CPR numbers were stolen at mass scale, the illegal access ran for around ten days, reporting from the 6 October press briefing puts the company account above 14 million lookup attempts, the minister says the security was not good enough, and an abnormally large bill exposed the abuse. On 7 October DI Digital stated that neither automatic alarms nor spot checks exposed it. What is still hidden is the trail that should explain how this kept running: the product, exact request count, timing curve, customer baseline, first anomaly, alert history, credential path and the control that failed.

01

ABOUT 8.8 MILLION PEOPLE HAD THEIR NAMES, ADDRESSES AND CPR NUMBERS STOLEN.

CPR's own notice says the illegal access reached names, addresses, CPR numbers and other information concerning about 8.8 million registered people, yet the public still has no incident-specific way to check person by person who was hit.

02

THE ILLEGAL ACCESS KEPT RUNNING FOR AROUND TEN DAYS.

The minister told Ritzau the activity ran for around ten days in September, while CPR says irregular behavior was first noticed on the evening of 2 October; that ten-day window is what the technical record now has to explain.

03

THE AUTHORITIES KNOW THE COMPANY AND REFUSE TO NAME IT.

Ritzau reported on 6 October that the authorities will not say which company it is even though they know its identity, and the CPR product plus credential path used in the illegal access remain withheld as well.

04

MORE THAN 14 MILLION CPR LOOKUP ATTEMPTS ARE NOW REPORTED.

Reporting from the 6 October press briefing says the company account made more than 14 million lookup attempts, with about 8.8 million returning a record. That is no longer an abstract "very large number"; it is a public floor above fourteen million attempts, while the exact count and request curve remain hidden.

05

THE MINISTER ADMITS THE SECURITY AROUND THE ACCESS WAS NOT GOOD ENOUGH.

Christina Egelund said the security around this access was not good enough and warning lights should have gone off earlier, so this failure comes from the minister's own words, not from me reading between the lines.

06

THE BILL EXPOSED THE ABUSE; AUTOMATIC ALARMS AND SPOT CHECKS DID NOT.

On 7 October DI Digital, citing information confirmed at the 6 October press briefing, states that neither automatic alarms nor spot checks exposed the abuse. The unusually large September bill did. That turns the missing alert history from a fair question into a central record they need to publish.

AFTER AROUND TEN DAYS OF ILLEGAL ACCESS, A BILL EXPOSED THE SCALE BEFORE THE PUBLIC WAS TOLD.

SEPTEMBER

Illegal automated CPR activity runs for around ten days, according to the minister.

02 OCT

CPR administration notices abnormal activity. A later ministry briefing says an abnormally large bill exposed the scale.

WEEKEND

CPR says the scale becomes clear while the incident is mapped.

04 OCT

CPR notifies Datatilsynet about the breach.

05 OCT

The data theft is announced publicly. Police investigation and a security review are underway.

05 OCT / 11:57

CPR states that private companies' access to CPR has been restored.

NO INVENTED FIELDS, NO SOFTENED FACTS.
The official notice lists names, addresses, CPR numbers and “other information” without publishing the complete field list. I will not invent fields the record does not show or soften the ones it does: names, addresses and CPR numbers were stolen.
02 / THEIR OWN CONTROL TRAIL

YOU LOG THE ACCESS, COUNT THE USE AND REQUIRE REVIEW, SO SHOW THE TRAIL.

CPR's own terms require logs, transaction statistics, named security responsibility and recurring review for private access, records that should reconstruct where the abuse ran, how long it ran and who was tied to it. The authorities still have not disclosed the product at the center of the breach, so name the product and publish the trail that applies to it.

LOGGEDSEARCHES + QUERIES
COUNTEDTRANSACTIONS + USAGE
REVIEWEDSECURITY REVIEW
THE BILL EXPOSED THE ABUSE; AUTOMATIC ALARMS AND SPOT CHECKS DID NOT. THEIR OWN RULES STILL REQUIRE LOGS, COUNTERS AND REVIEW, SO SHOW WHAT THOSE CONTROLS SAW DURING THE TEN DAYS AND WHY NONE STOPPED THE ACCESS.
CTRL 01
THEIR RULEIF DATAFORDELER WAS THE ROUTE, SHOW THE LOGPOINT TRAIL.

CPR's private Datafordeler terms say searches and queries are recorded with the client user or certificate, transaction type, date, time and queried data, while designated security users can report selected or all searches and receive transaction statistics.

RECORD THAT SHOULD EXIST

Logpoint export covering user or certificate, transaction type, date and time, queried data, search reports and transaction statistics.

PUBLIC GAP

If this was the route, publish the incident-period Logpoint trail, alert history and security-control record.

CTRL 02
THEIR RULEIF SYSTEM-TO-SYSTEM WAS USED, THE MONTHLY STATISTICS AND SECURITY REVIEW SHOULD EXIST.

For private system-to-system deliveries outside Datafordeler, CPR requires a named security officer, monthly transaction statistics by type and review by that officer, and it says online batches need prior CPR approval.

RECORD THAT SHOULD EXIST

September transaction statistics, security-officer review, client-side employee attribution and any CPR approval for online batching.

PUBLIC GAP

If this was the route, publish the statistics, user mapping and any approval tied to the activity.

CTRL 03
THEIR RULEIF CPRWEB WAS USED, EVERY QUERY SHOULD HAVE A USER, TIME AND TARGET.

Private CPRWeb terms say every query is logged with user ID, date and time, and the query target, while the security officer reviews each user's transaction statistics monthly and can request detailed terminal traffic for a defined period.

RECORD THAT SHOULD EXIST

Per-query user ID, timestamp and query target, monthly per-user transaction statistics and detailed terminal traffic for the relevant users and period.

PUBLIC GAP

If this was the route, publish the suspected-abuse query report and the September user statistics.

CTRL 04
THEIR RULETHE CUSTOMER WAS IDENTIFIED BEFORE ACCESS WAS GRANTED.

CPR's application process asks for the applicant's CVR, data controller, named security officer, billing contact and requested products, so the company may be hidden from the public now, but it was not an anonymous customer inside the access model.

RECORD THAT SHOULD EXIST

Customer application, CVR, approved purpose, requested product, named security officer and responsible contacts.

PUBLIC GAP

The authorities still withhold the company name, product and accountable security contact for the access path at the center of the breach.

CTRL 05
THEIR RULETHE PLATFORM ALREADY COUNTS USAGE, AND THE BILL PROVED THAT MATTERS.

Published pricing already counts CPRWeb searches, CPR Direkte lookups, CPR Services transactions and Datafordeler usage; REST counts even zero-result transactions, while GraphQL pricing counts returned person objects, so those counters should be checked against the incident telemetry.

RECORD THAT SHOULD EXIST

Usage counters and billing records for searches, lookups, REST transactions or returned GraphQL person objects, depending on product.

PUBLIC GAP

Publish the customer baseline, September usage curve and the exact point where volume broke from normal behavior.

CTRL 06
THEIR RULENAME THE PRODUCT AND SHOW THE TRAIL THAT GOES WITH IT.

The route decides what evidence should exist: Datafordeler means Logpoint and transaction statistics, CPRWeb means query logs and per-user statistics, and another system-to-system product means monthly statistics plus system and client-side user attribution.

RECORD THAT SHOULD EXIST

Product identity, transaction semantics, limits, approvals, IP or certificate history and the exact control path that applied during the incident.

PUBLIC GAP

Without the product, claims about the failed control stay incomplete, so name the route and show its trail.

I am not asking CPR to create evidence after the fact. Their own terms already describe records that should exist, and the 7 October reporting makes the gap sharper: publish the product, exact request count, usage curve, first anomaly, full alert history, spot-check record, credential path, user attribution and every control change completed before private access was restored.
PRECEDENT, NOT THIS INCIDENT A PRIVATE CPR ACCESS PATH HAD ALREADY BEEN ABUSED BEFORE THIS CASE.

DI Digital points to a separate December 2025 case in which demo accounts at a payroll platform were used for nearly 90,000 unauthorized CPR lookups. It is not evidence of the same company or route here. It is evidence that abuse of an intermediary CPR access path was not a hypothetical risk before September 2026.

03 / SCALE

MORE THAN 14 MILLION LOOKUP ATTEMPTS ARE NOW REPORTED, WHILE THE EXACT COUNT, PEAK AND CUSTOMER BASELINE ARE STILL NOT PUBLIC.

The public record has moved. Reporting from the 6 October press briefing now puts the company account above 14 million CPR lookup attempts, with about 8.8 million returning a record. The exact count, timestamps, peak rate, customer baseline and CPR product are still withheld, but the scale is no longer just an inference from the number of people hit.

MINIMUM AVERAGE FROM THE REPORTED 14M+ ATTEMPTS, NOT THE REAL REQUEST RATEIf the reported 14M+ attempts are spread evenly across a full ten days, the minimum average is already:
>1.4Mlookup attempts / day
>58,333lookup attempts / hour
>972lookup attempts / minute
>16.2lookup attempts / second
>14,000,000 / (10 x 24 x 60 x 60) = >16.20 attempted lookups per second on a flat ten-day average

This is not the real request rate and it is not a peak. Public reporting gives only a floor above 14 million attempts and an approximate ten-day window. Bursts, failed attempts and downtime can move the real curve sharply, so publish the exact count, timestamps, peak, product and customer baseline.

14M+ ATTEMPTS, ABOUT 8.8 MILLION PEOPLE HIT, AROUND TEN DAYS, AND THE BILL CAUGHT IT. DI Digital now says automatic alarms and spot checks did not expose the abuse. Publish what monitoring actually saw, when it saw it and why it did not stop the access.
04 / ACCESS MODEL

NAME THE PRODUCT, BECAUSE WITHOUT IT NO ONE CAN MATCH THIS THEFT TO THE LOGS, LIMITS AND REVIEWS THAT APPLIED.

CPR documents several private routes into the register: web searches, program-to-program access, services, Datafordeler APIs and file extracts. Those routes do not leave the same technical trail. Until the authorities say which product the company whose access was abused used, the public cannot know which credentials, counters, approvals, logs and review duties should be checked against around ten days of illegal access.

CPRWEB

Private companies can search CPRWeb with identifying information and receive current name, address and other permitted data. If this was the route, the per-query user, time and target logs matter immediately.

CPR DIREKTE

CPR Direkte is program-to-program access from a customer system into CPR. If this route was abused, the system-to-system authorization, employee attribution and transaction trail are central.

CPR SERVICES

CPR Services lets customer systems make structured requests against CPR services. If this was the route, the transaction semantics, customer authorization and usage records belong in the incident trail.

DATAFORDELER

Datafordeler exposes CPR person data through REST and GraphQL. CPR's own terms describe Logpoint records and transaction statistics for this route, so if it was used, that trail should exist.

EXTRACTS AND FTP

CPR also supports status and change extracts, including recurring flat-file deliveries for defined populations. If an extract route was involved, the product and delivery trail need to be identified and published.

THEY REFUSE TO NAME THE COMPANY, AND THEY STILL HAVE NOT SAID WHICH PRODUCT WAS USED.

This omission blocks independent verification. Name the product, credential type and query semantics, then publish the matching logs, counters, approvals, alert history and review record. Without the route, nobody outside the system can test the official explanation against CPR's own rules.

CPR ALREADY SUPPORTS LARGE-POPULATION WORK, SO RAW VOLUME ALONE CANNOT SEPARATE LEGITIMATE USE FROM ABUSE.

Adressematch documentation recommends a match analysis on up to 5,000 people before matching a large population. That makes the control problem harder, not smaller: product, approved purpose, customer baseline, request pattern, velocity and authorization have to be read together. A system that legitimately handles bulk data needs controls that can tell bulk work from theft.

05 / THE DATA IS ALREADY OUT

CLOSING THE ACCESS DOES NOT PULL THE STOLEN DATA BACK.

Criminals already copied names, addresses, CPR numbers and other information, and shutting one route only stops more access through that route; it does not erase what was stolen. That is why, on 6 October, the Danish Agency for Societal Security told authorities, organizations and companies not to rely on CPR data alone for identity verification, pushing the consequences downstream onto everyone who now has to treat those identity details as no longer secret.

NAMES
ADDRESSES
CPR NUMBERS
OTHER INFORMATION
07 OCT / DATATILSYNET

SOMEONE KNOWING YOUR NAME, ADDRESS AND CPR NUMBER PROVES NOTHING ABOUT WHETHER THEY ARE LEGITIMATE.

On 7 October Datatilsynet said the stolen personal information can make phishing and smishing more convincing and that a CPR number should not itself be treated as a secret authentication factor. If a service uses a CPR number, or part of it, as a username or password, Datatilsynet says those login details should be changed. That is not a hypothetical risk note; it is the regulator speaking after the breach.

PRIVATE-COMPANY ACCESS WAS RESTORED AT 11:57, BUT THE PUBLIC STILL HAS NOT BEEN SHOWN WHAT SECURITY CHANGED FIRST.

CPR says private-company access was restored at 11:57 on 5 October; the material reviewed for this case still does not identify the monitoring, throttling, suspension or access-control changes completed before restoration, so publish the changes, the tests and the timestamps.

THE MINISTER WOULD NOT RULE OUT NEW CPR NUMBERS.

After briefing parliamentary committees on 5 October, Christina Egelund said it was too early to know whether some people might need new CPR numbers. No replacement decision has been announced, but once replacing national identity numbers is even being considered, the aftermath is plainly bigger than closing one abused access route.

DIRECT CONSEQUENCE / 06 OCTTHE STATE NOW TELLS AUTHORITIES AND COMPANIES CPR DATA ALONE IS NOT ENOUGH TO VERIFY WHO YOU ARE.

On 6 October the Danish Agency for Societal Security told authorities, organizations and companies to strengthen identity checks and use more than CPR data. That is an operational consequence at national level: the theft has already changed how identity has to be verified.

WHO CARRIES THE AFTERMATH

THE DATA WAS STOLEN, AND THE TIME, EXTRA CHECKS AND FRAUD RISK NOW LAND ON THE PUBLIC.

People are now expected to spend more time proving who they are, distrust calls and messages that know their identity details, watch for fraud and report it, and absorb disruption created by a failure they did not cause. At the same time, authorities still refuse to identify the company and affected people still have no incident-specific person-by-person answer about their own record.

EXTRA IDENTITY CHECKS
FRAUD WATCH + REPORTING
TIME LOST CLEANING UP
CLOSING ACCESS DOES NOT ERASE STOLEN DATA.
06 / WHO EXACTLY WAS HIT

8.8 MILLION IS ONLY THE TOTAL, AND PEOPLE STILL CANNOT CHECK WHETHER THEIR OWN CPR RECORD WAS AMONG THE STOLEN DATA.

As of 6 October, the official pages still give one aggregate number and general warnings, but no public incident check lets a person see whether their own CPR record was hit, which fields were stolen, when it happened or through which product. Ordinary CPR register access shows what CPR stores about you, not whether your record was part of this theft.

WHAT IS PUBLIC

WHAT THE 8.8 MILLION ACTUALLY COVERS

The total includes living people, people who moved abroad and deceased people. CPR also says names and addresses belonging to people with name and address protection were excluded from the unauthorized access. That narrows one part of the dataset, but it still does not tell an individual what happened to their own record.

WHAT IS MISSING

PEOPLE STILL CANNOT CHECK WHETHER THEIR OWN CPR DATA WAS STOLEN

There is still no public person-by-person answer showing whether a specific record was queried, which fields were stolen, when or how often it was queried, or which product was used. Those are basic questions when the data belongs to you.

THE TRAIL

THEIR OWN RULES DESCRIBE RECORDS THAT COULD ANSWER THIS

Depending on the product, CPR's own terms describe per-query logs, Logpoint records, monthly transaction statistics and user attribution. Name the product, preserve the relevant trail and tell people what happened to their own CPR data.

07 / SUPERVISION

DATATILSYNET SUPERVISES DATA PROTECTION, SO SHOW WHAT WAS CHECKED BEFORE THE THEFT REACHED 8.8 MILLION PEOPLE.

Datatilsynet says it received the breach notification on 4 October and is now examining what happened, how it could happen and who is responsible; what the public still cannot verify is the supervision before the theft, including which inspections or other supervisory actions covered private-company CPR access, mass automated lookups, access review, anomaly detection or the audit trails CPR's own terms say should exist.

MANDATE

SUPERVISION HAS ITS OWN RECORD.

Datatilsynet does not operate CPR; it supervises compliance with data-protection rules and carries out inspections of public authorities and companies. The operational breach record belongs to CPR, while any inspections, evidence requests, findings and follow-up from the supervisor form a separate accountability record that should also be made public.

BEFORE

SHOW WHAT WAS CHECKED BEFORE SEPTEMBER.

Publish any pre-incident inspection, audit or supervisory action that touched private-company CPR access, automated high-volume querying, anomaly detection, access review or the logs and transaction statistics required by CPR's own terms. If none covered those controls, say that plainly.

NOW

THE INVESTIGATION NEEDS VERIFIABLE FINDINGS.

Datatilsynet says it is examining what happened, how it could happen and who is responsible. The result should identify responsibility, control failures, any orders or sanctions, corrective actions and the evidence that proves those fixes were actually made.

ADVICE AFTER THE THEFT CANNOT REPLACE THE SUPERVISION RECORD FROM BEFORE IT.
Publish the supervision timeline: what was checked before September, what records are being demanded now, what the investigation finds and what enforcement follows. If no prior supervision covered these controls, publish that too.
08 / GDPR + LONG-TERM RISK

GDPR DOES NOT BECOME OPTIONAL WHEN PUBLIC INFRASTRUCTURE FAILS.

Datatilsynet's own guidance says breach risk must include future misuse after data leaves the controller's control, and it warns that even ordinary personal data can be used for targeted phishing and fraud. GDPR therefore leaves concrete questions here: was the supervisory authority told in time, does the breach create high risk for the people hit, and if so how are those people being told what happened to them? General warnings do not answer those questions person by person.

GDPR 01

ARTICLE 33: THE PUBLIC DATES DO NOT SHOW A MISSED 72-HOUR DEADLINE

CPR says it noticed irregular activity on 2 October, and Datatilsynet says the breach notification arrived on 4 October. On the dates now public, I cannot honestly claim the 72-hour deadline was missed. If an earlier awareness timestamp exists, publish it, because Article 33 runs from when the controller becomes aware of the breach.

GDPR 02

ARTICLE 34: HIGH RISK CAN REQUIRE DIRECT NOTICE TO THE PEOPLE HIT

When a breach is likely to create high risk, GDPR requires communication to the affected people without undue delay unless an Article 34 exception applies. If the controller relies on an exception or on public communication instead of individual notice, name the legal basis and show why the chosen method informs the people hit just as effectively.

GDPR 03

8.8 MILLION IS NOT A PERSON-BY-PERSON ANSWER

An aggregate number does not tell me or you whether a specific CPR record was queried, what fields were stolen, when it happened or how often it was touched. As of 6 October, the public material reviewed here still provides no incident-specific self-check that answers those questions for an individual.

GDPR 04

NAME-AND-ADDRESS PROTECTION STILL LEAVES A CPR-NUMBER QUESTION

CPR says names and addresses belonging to people with name and address protection were excluded from the unauthorized access. The public announcement still does not clearly say whether CPR numbers for that protected group were accessed, and that distinction should be answered directly.

GDPR 05

REVOKING ACCESS DOES NOT ERASE STOLEN DATA.

Revoking the company's access can stop one route, but it cannot erase copies already stolen, identify every downstream copy or remove future fraud risk. Datatilsynet's own guidance says breach risk can extend years into the future once data has left the controller's control.

09 / AUTHORITIES REFUSE TO NAME THE COMPANY

THE AUTHORITIES KNOW WHICH COMPANY AND REFUSE TO TELL THE PUBLIC.

Ritzau reported on 6 October that the authorities know the small Danish company whose authorized CPR access was abused and refuse to identify it publicly. At the same time, citizens and businesses are being told to tighten identity checks because stolen CPR data can support fraud, while the company name, product, credential path and full incident trail remain out of public view.

WHAT THE AUTHORITIES HAVE ADMITTED

A small Danish company had authorized CPR access, criminals abused that access for around ten days, and an abnormally large bill exposed the scale. The authorities know the company's identity and refuse to publish it.

WHAT THE PUBLIC STILL HAS NOT BEEN SHOWN

The public still does not have the company name, CPR product, credential type, compromise mechanism, exact request count, peak rate, normal baseline, alert thresholds, first anomaly timestamp, full person-by-person impact map or the destination of the stolen data.

THE PUBLIC CARRIES THE EXTRA RISK WHILE AUTHORITIES KEEP THE COMPANY NAME OUT OF PUBLIC VIEW AND THE PRODUCT, CREDENTIAL PATH, TELEMETRY, ALERT HISTORY AND PERSON-BY-PERSON IMPACT REMAIN UNPUBLISHED.
10 / QUESTIONS THAT REQUIRE RECORDS, NOT PR

ADVICE WILL NOT EXPLAIN TEN DAYS OF ILLEGAL ACCESS, SO SHOW THE LOGS, ALERTS, BILLING CURVE AND DECISIONS.

Telling people to watch for fraud does not explain why the illegal access ran for days, which CPR product was used, what monitoring saw, whose data was stolen or what changed before private access was restored. Those answers should already exist in logs, billing records, access-control configuration, contracts, incident tickets and formal risk records.

Which private company was it, and on what basis are authorities withholding its identity from the public?
Which CPR product and interface were used for the illegal activity?
Which credential or identity was used, and was it compromised, stolen or misused?
What were the exact start and end timestamps, and when did the first anomaly appear?
Reporting now puts the company account above 14 million lookup attempts. What is the exact total, including failed and zero-result lookups?
How many unique people were returned, and which fields were returned through each product path?
What was this customer's normal request baseline, what peak was reached, and how far did the activity deviate from normal?
Which rate limits, quotas, anti-enumeration controls and automatic suspension rules existed on the route that was abused?
DI Digital says automatic alarms and spot checks did not expose the abuse. What did those controls show during the ten days, which alerts fired, who saw them and why did none stop the access?
Why did a bill expose the abuse before the security controls did, and which control was supposed to catch this volume first?
Can every affected person obtain an incident-specific record showing whether their CPR number was queried and what data was returned?
Were CPR numbers belonging to people with name and address protection accessed even though their protected names and addresses were excluded?
What legal and risk assessment supports the chosen method of informing affected people under GDPR Article 34?
What Datatilsynet inspections, audits or supervisory actions before the breach covered private-company CPR access, high-volume automated lookups, anomaly detection or the audit trails CPR itself requires?
Which concrete controls were changed, tested and approved before private-company CPR access was restored at 11:57 on 5 October?
PUBLIC RECORD STILL INCOMPLETE
11 / PRESERVED EVIDENCE / PROOF FIRST / PROOF FIRST.

SOURCE, TIMESTAMP, HASH, LIMIT.

The public evidence is deliberately narrow. It keeps the official incident record, the control trail, the written findings and the hashes needed to verify what I published, while leaving out crawl maps, error-path inventories, security-header matrices and preproduction or password-workflow details that do not establish the breach path and would add attack-relevant detail without strengthening the case.

VERIFY THE CASE, NOT THE TARGET.The public package contains no credentials, sessions, personal CPR records, authenticated material, exploit material, raw crawl map, error-path inventory or security-header matrix. The evidence that matters for accountability stays public; operational detail that only improves reconnaissance stays out.
SHA-256 verifies file identity, not causation. A hash can prove that a file did not change; it cannot prove why the theft happened. Every claim still stands or falls on its source and its written limit.
12 / SOURCES

THE RECORD STARTS WITH THEIR OWN WORDS AND THEIR OWN PUBLIC DOCUMENTATION.

These are the documents the case stands on: CPR, the ministry, Datatilsynet, Ritzau, CPR's own access terms and the operational notices issued after the theft. The 6 and 7 October record adds three hard facts: more than 14 million lookup attempts are now reported, automatic alarms and spot checks did not expose the abuse, and Datatilsynet says a CPR number should not itself be treated as a secret authentication factor. The company, product, exact request count, peak, full alert history and control changes remain withheld.

CPR: extensive unauthorized access to citizen CPR information 05 OCT 2026 / OFFICIAL CPR INCIDENT NOTICE Ministry: extensive unauthorized access to CPR information 05 OCT 2026 / OFFICIAL MINISTRY FACTS + TIMELINE Datatilsynet: case concerning CPR lookups 05 OCT 2026 / OFFICIAL SUPERVISORY NOTICE + AUTOMATED LOOKUPS Ritzau via DK Nyt: misuse continued around ten days 05 OCT 2026 / RITZAU REPORTING / TEN DAYS + INADEQUATE SECURITY CPR: description of products CPR DOCUMENTATION / PRODUCT + ACCESS MODEL Datatilsynet: risk assessment guidance DATATILSYNET GUIDANCE / LONG-TERM MISUSE RISK Datatilsynet: serious criticism in notification case DATATILSYNET DECISION / HIGH-RISK NOTIFICATION CONTEXT CPR standard terms for private data deliveries, 16 June 2026 16 JUN 2026 / CPR TERMS / SYSTEM-TO-SYSTEM CONTROLS CPR standard terms for private CPRWeb access, 16 June 2026 16 JUN 2026 / CPR TERMS / CPRWEB QUERY LOGGING + REVIEW CPR standard terms for private Datafordeler access, 16 June 2026 16 JUN 2026 / CPR TERMS / DATAFORDELER LOGPOINT + CONTROL CPR: private-company access restored at 11:57 on 5 October 05 OCT 2026 / CPR NOTICE / PRIVATE ACCESS RESTORED 11:57 Ritzau / NI: CPR leak discovered because of an unusually large bill 06 OCT 2026 / RITZAU VIA NI / BILL DISCOVERY + COMPANY WITHHELD Danish Agency for Societal Security: strengthen identity control after CPR leak 06 OCT 2026 / NATIONAL IDENTITY-CONTROL CONSEQUENCE Borger.dk: current CPR incident warning and ordinary CPR register access 06 OCT 2026 / GENERAL WARNING / NOT AN INCIDENT SELF-CHECK Datatilsynet: organisation and supervisory role OFFICIAL / DATATILSYNET SUPERVISORY ROLE Datatilsynet: 2024 decision concerning access to CPR security logs 2024 / DATATILSYNET DECISION / CPR SECURITY-LOG ACCESS CONTEXT The Next Web: more than 14 million CPR lookup attempts through the company account 06 OCT 2026 / SECONDARY REPORT / 14M+ ATTEMPTS + ABOUT 8.8M RETURNED RECORDS DI Digital: automatic alarms and spot checks did not expose the abuse 07 OCT 2026 / INDUSTRY ASSOCIATION / BILL DISCOVERY + 2025 PRECEDENT Datatilsynet: what the CPR breach means for people now 07 OCT 2026 / OFFICIAL / PHISHING RISK + CPR IS NOT A SECRET AUTHENTICATION FACTOR
13 / EDITORIAL STANDARD

I DO THIS ALONE, BUT THE EVIDENCE STILL HAS TO SURVIVE CONTACT WITH THE OTHER SIDE.

I live in Denmark and my own CPR sits inside the same national identity system. I make my living outside IT, pay for my own tools and investigate independently, without ministry staff, a press office or a consultancy budget behind me. A title, position or degree does not turn a failed control into a working one. Show the record.

I stopped at public, unauthenticated material. I did not test live identity paths, submit forms, enumerate CPR numbers, bypass access controls or touch private data. Operational security observations that do not establish the breach path stay out of the public evidence package. Removing attack detail that does not prove the case leaves the evidence cleaner and harder to dismiss.

If a point here is wrong, publish the record that disproves it. If the logs, counters and reviews worked as designed, show them. If they did not, say exactly where the chain failed and what changed before private-company CPR access was restored.
OPEN CASE / UPDATED AS EVIDENCE CHANGES

THE DATA WAS STOLEN, THE BILL EXPOSED THE ABUSE, AND NOW THE FULL TECHNICAL RECORD HAS TO BE SHOWN.

The public now has a floor above 14 million lookup attempts, the minister has already said the security around the access was not good enough, and 7 October reporting says automatic alarms and spot checks did not expose the abuse. The record still needs to identify the company, CPR product and credential or identity used, including whether the access came from compromise, theft or misuse, and it needs to publish the exact request count, start and end timestamps, peak rate against the customer’s normal baseline, billing curve, first anomaly and full alert history. It also needs to show the rate limits, anti-enumeration controls and automatic suspension rules on the abused route, who owned monitoring, how the activity was attributed to a user or credential, and the person-by-person affected map together with the fields returned. Finally, publish the Article 34 notification basis, what prior supervision actually tested, every control changed before access was restored and where final accountability lands, because those are the records this case still lacks, not talking points.

THE DATA WAS STOLEN, AND THE FULL TECHNICAL RECORD IS STILL NOT PUBLIC.

TRACE