Written by: Mio Kanzaki(Security Analyst (AI))Reviewed by: 千葉圭介Updated
Unauthorized access prevention means stopping logins made with someone else's ID and password, and break-ins through system holes, from succeeding. If one does succeed, you want to notice it early. It concerns sites with admin screens such as WordPress and EC-CUBE, companies with member pages or internal systems, and the agencies that run those sites. Three jobs come first: strengthen logins (tidy passwords and add two-factor authentication), close entrances you do not use, and keep software updated.
This page is a hub. It helps a site owner find the entrances, decide which to close first and decide what to hand to a specialist. If you have already found a login or a transfer you do not recognize, read the emergency guide first.
What does unauthorized access mean in practice?
Unauthorized access means entering a system whose access is restricted without permission. In Japan the Unauthorized Access Prohibition Act forbids it (glossary entry). There are two ways in. One is to type someone else's ID and password. The other is to use a program flaw or a configuration mistake.
The first way, walking in through the front door, is easy to overlook. The intruder opens the door with a key they obtained. The key may be information stolen with phishing (a fake site that looks real and makes you type your details), a reused password, or an account left behind by a former employee or contractor.
What follows depends on the service that was entered: an unauthorized bank transfer, trades in a securities account, purchases at an online shop, reading of email, or a rewritten site. For that reason, ask "which entrance needs a lock?" as well as "what should we protect?".
Which entrances do intruders use?
The table lists the entrances that a company with a site or internal systems should inspect, arranged by the kind of key that opens them.
| Entrance | What happens | What protects it | Read more |
|---|---|---|---|
| Login to an admin area (WordPress, EC-CUBE and others) | Guessing or reuse gets someone in as an administrator | Two-factor authentication, limits on attempts, limits on where it can be opened from | Attacks on the WordPress admin screen |
| Brute-force login attempts | A machine tries huge numbers of combinations and gets in when one matches | Rate limiting (a setting that caps attempts in a short time) | Brute-force attacks |
| Reused passwords | A combination leaked from another service is used | Remove duplicate passwords and add two-factor authentication | Password reuse |
| Accounts without multi-factor authentication | The password alone opens the account once it leaks | Introduce it on all administrator, contractor and cloud accounts | Missing multi-factor authentication |
| Devices left on default settings | The product's standard password gets someone in | Change it at installation and do not expose unneeded devices | Default passwords left in use |
| VPN appliances and remote desktop | Someone from outside gets into the internal network | Update the device, narrow what is exposed, strengthen authentication | VPN appliance vulnerabilities, Exposed remote desktop |
| Mistakes in access rights | After login, a user can see another person's information | Check permissions screen by screen and data by data | Broken access control |
| Accounts of contractors and former employees | The key still works after the contract or employment has ended | Delete the ID and change passwords at the end | Intrusion through contractors |
The first four rows are the entrances to check first on this page. Each one is opened with a key, such as a password or an account, so key management is where the work starts.
We do not publish break-in steps or attack strings.
What do public statistics show?
Statistics let you compare your company with the national picture. The figures below come from the original releases, which we opened and read. A recognized case is one that police confirmed, so it is not the total number of actual victims.
Police, ministries: recognized cases rose 34.2% to 7,190
Under Article 10 of the Unauthorized Access Prohibition Act, the National Public Safety Commission, the Minister for Internal Affairs and Communications and the Minister of Economy, Trade and Industry publish this situation every year. In the release of 12 March 2026, recognized cases in 2025 (Reiwa 7) were 7,190. That is 1,832 more than the 5,358 of the year before, an increase of about 34.2%. Compared with the 1,516 cases of 2021 (Reiwa 3), it is about 4.7 times.
Among acts carried out after the unauthorized access, unauthorized transfers on internet banking and similar were the most numerous at 4,747. Next came unauthorized trades on securities companies' online services at 1,484, and unauthorized purchases on internet shopping at 228. The securities figure was only 2 the year before. The targets moved a long way in about a year.
For a site owner, the figure for website tampering or deletion matters. It was 13 in 2025, up from 9 the year before. It counts only what police recognized as unauthorized access. Tampering has its own hub, preventing website tampering.
What arrests show about methods
Arrests numbered 431 cases and 248 people. Of these, the type that enters by typing another person's ID and password (called the identification code theft type) accounted for 399 cases. Arrests for the act of unauthorized access itself numbered 407, so that type is more than 90% of them. The type that attacks a system defect (the security hole attack type) accounted for 8.
For those 399 cases, the sources of the IDs and passwords were as follows.
- Obtained by exploiting weak password setting or management: 84 cases (174 the year before)
- Obtained from phishing sites: 77 cases (41 the year before)
Cases from weak passwords fell to less than half, and phishing, which gets the owner to type the password, almost doubled.
The breakdown of services misused had dedicated sites for employees or members at 122 cases (30.6%), the most. Community sites followed with 92 (23.1%), internet shopping with 67 (16.8%) and internet banking with 46 (11.5%).
The cases the authorities published also show how entry happens. A group of men, occupation unknown, used generative AI in part to build and publish a phishing site imitating a large e-commerce site, and were arrested in June 2025. In another case, three teenagers used other people's IDs and passwords obtained improperly to enter a telecom carrier's server and obtain eSIMs improperly.
IPA: exploiting vulnerabilities ranks fourth, attacks on remote environments eighth
IPA's Top 10 Information Security Threats 2026 was released on 29 January 2026. For organizations, first place went to damage from ransomware attacks, second to attacks aimed at the supply chain or contractors, and third to cyber risk from the use of AI, which entered the list for the first time. Fourth was attacks that exploit system vulnerabilities. Eighth was attacks aimed at remote work environments and mechanisms.
The individual-user list includes unauthorized login to internet services. That list is not ranked, but it concerns operators too, because when members are logged into without permission, your side takes the inquiries.
The publications ask operators for a consistent set of measures. The police and ministry material lists the following measures for administrators of access, meaning the people who decide who may use a system.
- Keep logs and watch for suspicious logins
- Set passwords that cannot be guessed easily
- Delete the IDs of former employees and contractors
- Update web systems and VPN devices
- Introduce two-factor authentication or passkeys (logging in with a fingerprint or face)
- Introduce sender domain authentication (DMARC, SPF, DKIM)
To see what an outsider can see of your own site, use the free URL check. It reads public information only and costs JPY 0.
What should you do first, by company size?
Begin with the step that fits your company.
A company of a few people up to about 20, with site updates outsourced
Check whose key works where. A former agency contact, an employee who left, or a freelancer you once hired may still have an account in the admin screen.
First, open the user list in the admin screen and pick out names that do not match a face. Then add two-factor authentication for every administrator you keep. If an agency handles updates, check in the contract or a maintenance email who issues and deletes accounts.
A company of a few dozen with an internal network and several sites
Entrances multiply outside the website, so inspect the windows that face the internet. Write down the VPN appliance, remote desktop, management ports and cloud consoles, each with the name of the person responsible.
Update the devices first. Then close windows you do not use and add two-factor authentication to those you keep. If no list exists, making one is the first measure.
A company with member pages or e-commerce that accepts logins from users
You now guard your users' keys as well as your own. If a member's password leaked from another service and is tried on your site, a login can succeed even though you did nothing wrong.
To stop machines that try brute-force or reused combinations, you need a mechanism that counts short-term login failures and blocks them. Also send login notification emails, offer two-factor authentication or passkeys, and check which data each member can see. An assessment that covers screens behind login finds permission design mistakes.
10 things to do today
These 10 come from our fix menu and close the entrances of unauthorized access. WordPress is the example, though other CMSs and in-house systems follow the same logic. A note on each item says whether the admin screen is enough or a specialist is the safer choice.
- Take stock of administrator accounts. In the user list, delete people who no longer use the site, people who left, and accounts of previous agencies. Keep the number of administrators to the minimum.
- Review passwords. Change administrator passwords to long ones that do not repeat across services.
- Turn on two-factor authentication. A leaked password alone no longer opens the door, because a confirmation on a phone is also required. An authenticator app is the easy way to start (two-factor authentication).
- Limit login attempts. Stop connections that fail many times in a short period, so that brute-force guessing is much less likely to succeed. A plugin setting can do this.
- Stop login names from being visible from outside. Fix the places, such as author displays, from which a login name could be guessed.
- Change the login URL. The entrance to the admin screen becomes harder to find, so mechanical attempts tend to miss.
- Restrict the admin screen to fixed locations. Allow only fixed IP addresses. It involves server settings, so ask us if you are not sure.
- Close entrances you do not use. Turn off XML-RPC and any other channel you have no use for.
- Update the core and plugins, and delete unused parts. Check VPN devices and internal servers on the same day. A break-in that uses a hole stops when the update is applied.
- Set up a WAF, a mechanism that stops bad traffic at the entrance. It stops suspicious login attempts and bad traffic before they reach the site, and it can serve as a stopgap until a fix is released (WAF).
Start with items 1 to 4. From item 6 onward, a wrong setting can lock you out too, so take a backup and keep a second logged-in screen open while you work. Completing the list cannot promise that nothing will happen. It reduces the chance of a break-in and helps you notice one sooner.
Short of time? Contact us and we will look at which of these 10 your login setup still lacks.
What should you hand to a specialist?
Ask two things before doing a job yourself: can a mistake be undone, and does the job need a look inside the system?
- Assessment that includes screens behind login. Permission design mistakes and weak management of sessions (the logged-in state) cannot be seen from outside. A WordPress/CMS site assessment is from JPY 49,800 (excl. tax, up to 30 pages). Web application assessment Lite is from JPY 98,000 (excl. tax, up to 10 screens). Screens behind login and paid plugins push the amount up.
- Hardening the admin screen in one go. The WordPress security hardening pack covers changing the login URL, restricting access and limiting attempts together, and it is from JPY 49,800 (excl. tax). We carry out the items you need from the fix menu.
- Inspection of VPN appliances, servers and networks. A mistake can stop business, so a server and network assessment (from JPY 49,800 excl. tax, up to 5 IPs) lists the open windows before they are fixed.
- Investigation when you suspect unauthorized login. The cause has to be examined with the logs kept. For an urgent case, a same-day start adds JPY 20,000 (excl. tax) to the usual work.
- Continuous watch. Checking every day for failed logins and new administrators stops the moment a person forgets. Guard, our AI monitoring for WordPress, is JPY 5,000 per month (excl. tax) for Lite. Guard Pro, which includes blocking of unauthorized logins and two-factor authentication, is JPY 15,000 per month (excl. tax). Maintenance Lite, which also handles updates, is JPY 9,800 per month (excl. tax).
Our work is to confirm and fix what an assessment finds, to investigate suspected intrusions and to keep watch. Before delivery, a PDF lists every change and you try the real screens. Once delivered, the first month is covered by free support.
We cannot promise to stop every attack, because methods change. An assessment shows whether currently known holes exist and says nothing about the future.
Which pages cover the individual holes?
The pages below cover holes behind unauthorized access. Each one explains the mechanism, the check and the fix.
Login and authentication
- Brute-force attacks
- Password reuse
- Missing multi-factor authentication
- Attacks on the WordPress admin screen
- Default passwords and settings left in use
Permissions and APIs
- Broken access control
- Broken object level authorization (BOLA/IDOR)
- Weak API authentication
- Missing rate limiting
Devices, cloud and contractors
- VPN appliance vulnerabilities
- Exposed remote desktop (RDP)
- Intrusion through contractors
- Excessive agency of AI agents (a hole where the permissions given to an AI are too broad and can be abused)
The other hubs are hacking prevention, preventing phishing and preventing website tampering.
How can the free URL check help you decide?
Choose what to close only after seeing your current position. Enter your site's address into the free URL check. It looks at public information only: certificate expiry, security headers, versions of public components and their known holes, and email authentication settings. It never attempts a login and sends no attack. It costs JPY 0.
If a result concerns you, we can sign a non-disclosure agreement (NDA), receive your assessment consent form and run the free assessment, which combines a quick tool scan with a person's review. The issue report comes first and the decision on fixes is yours.
Frequently asked questions
How is unauthorized access different from hacking or tampering?
Unauthorized access is the legal term for getting into a system without permission, either with someone else's ID and password or through a hole in the system. Stealing information, rewriting pages or moving money afterwards is the damage that follows. Closing the entrances is unauthorized access prevention, and preparing for rewrites is a separate job under tampering prevention.
Will a complex password stop unauthorized access?
A complex password alone is not enough. According to the police and ministries, in arrests for entering with another person's ID and password, 77 cases in 2025 involved credentials obtained from phishing sites, up from 41 the year before. Once the user hands the password over, its complexity no longer matters. Add a check that is not a password, such as two-factor authentication.
Failed login attempts show up in my logs every day. Can I ignore them?
If they only remain as attempts, nobody got in. It does mean the same door is being tried again and again. Limit login attempts and add two-factor authentication to make a successful login much less likely. If a successful login shows a time or place you do not recognize, contact us through the emergency guide.
How much does unauthorized access prevention cost?
The free URL check, which looks only at what is visible from outside, costs JPY 0. A WordPress/CMS site assessment is from JPY 49,800 (excl. tax, up to 30 pages). Page count, screens behind login and paid plugins all change the amount. Login hardening can be chosen from the fix menu inside the WordPress security hardening pack, from JPY 49,800 (excl. tax). A formal quotation follows the free assessment.
Next step
Start with items 1 to 4 above, and run the free URL check on your site today.
To discuss which entrances to close first, contact us. If you have already found a login or a transfer you do not recognize, use the emergency guide first.
Written by Mio Kanzaki (AI security analyst). Supervised by 千葉圭介. Updated 19 November 2026.
Frequently asked questions
How is unauthorized access different from hacking or tampering?
Unauthorized access is the legal term for getting into a system without permission, either with someone else's ID and password or through a hole in the system. Stealing information, rewriting pages or moving money afterwards is the damage that follows. Closing the entrances is unauthorized access prevention, and preparing for rewrites is a separate job under tampering prevention.
Will a complex password stop unauthorized access?
A complex password alone is not enough. According to the police and ministries, in arrests for entering with another person's ID and password, 77 cases in 2025 involved credentials obtained from phishing sites, up from 41 the year before. Once the user hands the password over, its complexity no longer matters. Add a check that is not a password, such as two-factor authentication.
Failed login attempts show up in my logs every day. Can I ignore them?
If they only remain as attempts, nobody got in. It does mean the same door is being tried again and again. Limit login attempts and add two-factor authentication to make a successful login much less likely. If a successful login shows a time or place you do not recognize, contact us through the emergency guide.
How much does unauthorized access prevention cost?
The free URL check, which looks only at what is visible from outside, costs JPY 0. A WordPress/CMS site assessment is from JPY 49,800 (excl. tax, up to 30 pages). Page count, screens behind login and paid plugins all change the amount. Login hardening can be chosen from the fix menu inside the WordPress security hardening pack, from JPY 49,800 (excl. tax). A formal quotation follows the free assessment.