July's big case: “WP2Shell” broke straight into WordPress core
On 17–18 July 2026, two related vulnerabilities were disclosed directly in WordPress core code. Not in a single plugin, as is usual, but in the platform itself, which runs on millions of sites. The impact was huge.
It was a chain of two vulnerabilities that the security industry started calling “WP2Shell”:
- CVE-2026-60137: an SQL injection vulnerability affecting WordPress versions 6.8.0–6.8.5, 6.9.0–6.9.4 and 7.0.0–7.0.1.
- CVE-2026-63030: a logic error in REST API “batch” handling in versions 6.9.0–6.9.4 and 7.0.0–7.0.1. It mixed up the validation and execution of requests, so requests were treated as trusted before they had been checked.
On their own these would have been hard to exploit. Combined, they allowed code execution without logging in. An attacker could create an administrator account and, for example, upload a malicious plugin. In practice that meant taking over the site completely without any authentication.
Severity ratings varied by source. WPScan rated it 9.8 on the CVSS scale (critical), while CISA's own rating was more moderate at 7.5 (high). [5] Either way, it was an unusually dangerous combination, because the attack required no login.
The WordPress community reacted quickly. The fixes (versions 6.9.5, 7.0.2 and 6.8.6) were released on 17 July, and they forced automatic updates on even for sites where these are normally switched off. [3] Even so, mass exploitation began over the same weekend. According to The Register, tens of thousands of attack attempts were documented, and different attackers managed to create more than a hundred backdoor accounts in the 24 hours after the fix was released. In some cases remote access malware was also installed on sites. Public exploit code appeared online within hours of disclosure, and the number of exploits quickly grew into dozens. [4]
The US cybersecurity agency CISA added both vulnerabilities to its list of known exploited vulnerabilities (the KEV catalogue) on 21 July. [2]
The Finnish perspective:
The National Cyber Security Centre Finland at Traficom published its own vulnerability advisory (opens in a new tab) on the case on 19 July and updated it on 23 July. The advisory states plainly that exploitation was observed widely “globally and in Finland”. So this was not something that only affected sites abroad. [1]
The centre's recommendations were in line with those of international bodies:
- update immediately to a fixed version
- check your logs for suspicious REST API requests and new administrator accounts
- if you suspect your site has been targeted, take disk and memory images for forensic analysis
Cybersecurity: The NIS2 Directive Takes Effect in Finland
Blog· 3 min read
July's plugin wave: more than 100 new vulnerabilities
Alongside the core vulnerability, July was a typically busy month on the plugin front too. According to Sucuri's monthly roundup, more than 100 new plugin vulnerabilities were published in July, affecting millions of WordPress installations in total. [6]
The most critical were SQL injections that required no login, including:
- Kirki theme framework (≤ 6.0.12, fixed in 6.0.13)
- Appointment Booking Plugin (≤ 5.6.3, fixed in 5.6.4)
- Events Manager (≤ 7.3.6, fixed in 7.3.7)
- Amelia booking plugin (≤ 2.4.2, fixed in 2.4.3)
- Bookly booking plugin (≤ 27.7, fixed in 27.8)
Among the high-severity issues, several different types stood out: a CSRF vulnerability in Loco Translate that led to code execution (≤ 2.8.5), unauthorised file reading in W3 Total Cache (≤ 2.9.4), account takeover through email header manipulation in Essential Addons for Elementor (≤ 6.6.10), and several XSS and privilege escalation vulnerabilities in the GiveWP donation plugin.
Another security company, Shield Security, also highlighted from the same period an SQL injection in the Ultimate Member plugin (severity 9.3/10, fixed in version 2.10.2), privilege escalation in SureCart (8.1/10), a full code execution vulnerability in WP Ultimate CSV Importer (10/10) and an administrator account takeover vulnerability in miniOrange's Social Login plugin (9.8/10). [7]
The most common single vulnerability type in the whole month's data was stored XSS (cross-site scripting). According to Sucuri, it affected more than 50 different plugins because of insufficient input sanitisation. It is a good reminder that many vulnerabilities still come from the same well-known types of error. [6]
At Into-Digital we do not use a single one of these plugins, so these plugin vulnerabilities were harmless for our clients.
A separate and still unresolved risk: miniOrange's SSO plugin
One of July's most worrying findings concerned miniOrange's “OAuth Single Sign-On” plugin (CVE-2026-57807, published 10 July 2026). [9] The problem was broken authentication logic in the password reset function. The vulnerability allowed an authentication bypass without logging in, which let an attacker pose as any user. Including an administrator.
Patchstack rated the severity at 9.8/10 (critical) and warned that the plugin is used on millions of sites up to version 38.5.8. [8] What made this case especially worrying was that no official fix was available at the time of disclosure. The recommended action was to deactivate the plugin completely or, alternatively, to use a “virtual patch” from a security service as temporary protection until the vendor releases a fix.
This is a good reminder that plugins related to login and SSO deserve special attention, because their vulnerabilities give direct access to managing the whole site.
Our clients do not use this plugin, so the problem does not affect them.
When the update itself is the malware
On 28 July a case came to light that broke the usual “a bug was found, fix it” pattern. Wordfence's AI-based threat detection system found that an update to the popular “Advanced Responsive Video Embedder” plugin (version 10.8.7) contained a backdoor embedded directly in the code (CVE-2026-18072, published 29 July 2026). This was a supply chain attack: the malicious code came through the official WordPress.org directory, not through a third-party site.
The malicious code worked through a file called “fn-update-check.php”, which ran early as WordPress started up. It compared a token sent by the attacker with a fixed SHA-256 hash coded into the plugin. If the token matched, it logged the attacker in as a randomly chosen administrator account, with no credentials at all. The whole attack needed only one HTTP request, and the code also sent site data to a server controlled by the attacker.
Around 20,000 sites were estimated to be at risk. Wordfence identified the infected version about two hours after it was released, and WordPress.org removed the plugin from distribution soon after. [10]
Site owners were advised to remove the plugin immediately, review user accounts, change authentication keys and invalidate active login sessions if the plugin had been in use during the critical window.
Our clients do not use this plugin, so the problem did not affect them.
August: WordPress 7.0.3 and 7.0.4 closed more holes
August continued with two significant core releases.
WordPress 7.0.3, released on 6 August 2026, fixed 12 vulnerabilities at once. [11] The most significant was a pre-authentication reflected XSS vulnerability on the login page (CVE-2026-64638 / GHSA-52p2-r8wf-jcrf), caused by insufficient sanitisation of the username field on failed login attempts. [12] Security researchers called the exploit chain “XSS2Shell”. The attacker injects malicious JavaScript into the login page. If an administrator opens a malicious link, the code can create Application Passwords through the REST API. These can then be used to upload a plugin containing a PHP web shell: code that runs even when the plugin is not activated.
It is worth being precise about severity. Some coverage called this “critical”, but the official CVSS 4.0 rating (HackerOne) is 8.9, in the “High” category, not “Critical”. [12] To succeed, the attack needs a targeted phishing message and a click from an administrator. It is not as straightforward as July's WP2Shell.
The other fixes in 7.0.3 covered four stored XSS vulnerabilities (in emoji settings, the “Post Content” block, the Quick Edit function and the “Post Date” block, all requiring at least a Contributor-level account), a privilege escalation in multisite environments that allowed unauthorised creation of sites, information leaks (such as exposed comments on password-protected posts and guessable post URL slugs), CSS injection through a bypass of the Safe CSS filter and a server-side request forgery (SSRF) in URL validation. An interesting detail: one of these vulnerabilities (the CSS injection) was reported to the WordPress security team by Anthropic. [11]
Only a week later, on 12 August, WordPress 7.0.4 was released. It fixed one more critical problem: a logged-in user at Author level could execute code through a file upload on sites that use the Imagick and Ghostscript libraries for image processing (CVE-2026-65640 / GHSA-8vr3-7mxf-gx8w). [13] This fix was backported unusually widely. Support updates went out to all maintained branches as far back as version 4.7, because the vulnerability affected several old core versions. WordPress 4.6 and older no longer receive security updates at all.
What do maintenance and protection for a web service cost?
The estimate takes a few minutes and does not ask for your contact details.
Why and how did the attacks happen?
The following section is interpretation and reasoned speculation about what drives attack waves like these. It is based on what published sources have observed, although attackers' real motives cannot be known for certain.
Why WordPress?
WordPress runs an estimated good third of all websites on the internet, which makes it an extremely attractive target for attackers. When the same core vulnerability affects millions of sites, one working exploit scales instantly to a huge number of targets. Most of the time this is not a targeted attack on a particular site but mass scanning. Automated bots combed the web around the clock looking for vulnerable versions. Any unpatched site, whether a small business, a blog or a webshop, became a target simply because it was vulnerable, not because it was especially interesting.
What motivates the attackers?
The reported attacks show a familiar range of ways to make money. Backdoor accounts and remote access malware (such as the Overlord RAT seen in the WP2Shell case) give a lasting foothold that can be sold on or turned into a botnet. Attackers often hide spam links and hidden redirects on the site (“SEO spam”) to feed off the site's search visibility. Other typical goals are cryptomining, stealing payment card or user data from webshops and setting up ransomware. A supply chain attack like the Video Embedder case is an especially efficient version of this. When malicious code is embedded in a popular plugin update, the attacker gets thousands of sites at once without having to break into each one separately.
How did an attack play out in practice?
The pattern repeated from case to case. First a researcher or vendor disclosed the vulnerability and the fix. Within hours, public proof-of-concept code appeared online and quickly turned into automated attack tools. The power of WP2Shell came specifically from chaining. A REST API logic error and an SQL injection, each hard to exploit alone, became full unauthenticated code execution when combined. Once they had a foothold, the attacker created an administrator account and uploaded a malicious plugin containing a web shell for persistent access. August's XSS2Shell was a subtler route that relied on human error. It needed an administrator to click a malicious link, after which creating application passwords opened the way to a web shell.
Why so fast?
The clearest common factor was that the window for exploitation collapsed. Several researchers suspected that AI-assisted tools sped up building working attack code from a disclosure. Historically this has taken a day or more, but in summer 2026 some exploits were ready within hours. This changes the arithmetic of defence fundamentally. “We'll update next week” is no longer a realistic plan for critical vulnerabilities.
What can we learn from this?
Three themes recur in all the cases from July and August 2026, and they are worth turning into concrete lessons.
First, the window for exploitation has narrowed dramatically. When a working attack can be ready within hours of disclosure, a site that updates a week later is practically defenceless against critical vulnerabilities.
The lesson: switch on automatic updates at least for security fixes, and make sure you have working backups so that updating does not feel risky.
Second, the risk is not limited to coding errors in third-party plugins. The Video Embedder case showed that deliberate malicious code can also spread through the official plugin directory. The miniOrange case showed that login-related plugins can stay unpatched for a long time.
The lesson: reduce your attack surface by removing plugins you do not use, prefer actively maintained and widely used plugins, and follow advisories for login and payment plugins especially closely.
Third, WordPress core itself is not immune. Two separate critical or high-severity core vulnerabilities within a couple of months are a reminder that following core updates matters as much as following plugin updates.
The lesson: do not assume that “once core is up to date, everything is fine”. Also keep an eye on attack surfaces such as the REST API and authentication, and consider a web application firewall (WAF) that offers virtual patching when a vendor's fix is delayed.
The practical checklist stayed the same throughout the period: keep core, themes and plugins up to date, preferably with automatic updates for critical fixes. Remove unused plugins, limit anonymous REST API access where needed, take regular backups, and follow both international security sources and the vulnerability advisories of the National Cyber Security Centre Finland.
Would you like us to look after your site's updates?
We explain how Care (Huolenpito) works and what it costs.
Mikko Erola
Account Director / Team Lead
Ask us more about maintaining and securing your web service. We will get back to you.
Frequently asked questions
Is my site at risk if I use WordPress?
If your WordPress core, theme and plugins are up to date (as they are for our clients), you are protected against these known vulnerabilities. The biggest risk is for sites that have not been updated or that have automatic updates switched off. The WP2Shell fix was shipped as a forced automatic update, so many sites were protected automatically back in July.
What is WP2Shell?
WP2Shell is the name the security industry gave to a chain of two WordPress core vulnerabilities (CVE-2026-60137 and CVE-2026-63030). Together they allowed code execution without logging in and a complete takeover of the site. They were fixed in WordPress versions 6.9.5, 7.0.2 and 6.8.6.
Which WordPress version should I update to?
As of August 2026, the latest maintained version (7.0.4 or newer) contained all of the summer's fixes. The general rule is to always update to the newest available version, because only the latest version is actively supported.
How do I check whether my site has already been compromised?
Check whether unknown administrator accounts, suspicious plugin files or unusual REST API requests have appeared in your logs. If you suspect a breach, contact a security professional and keep your logs and backups for the investigation.
We have applied all the latest WordPress security releases and the necessary plugin updates to the sites of all our Care clients. At the same time, we reviewed the state of the sites for any anomalies. Although we and our clients came through this well, we keep improving our update process so that security updates go live even faster and more reliably in future.
Sources:
- National Cyber Security Centre Finland (Traficom): WordPress-tuotteessa kriittinen haavoittuvuus (Critical vulnerability in WordPress). https://kyberturvallisuuskeskus.fi/fi/haavoittuvuudet/haavoittuvuus-2026-18 (opens in a new tab)
- Rapid7: CVE-2026-63030 (wp2shell): a Critical Remote Code Execution Vulnerability in WordPress Core. https://www.rapid7.com/blog/post/etr-cve-2026-63030-wp2shell-a-critical-remote-code-execution-vulnerability-in-wordpress-core/ (opens in a new tab)
- SecurityWeek: WP2Shell WordPress Vulnerabilities Exploited in the Wild. https://www.securityweek.com/wp2shell-wordpress-vulnerabilities-exploited-in-the-wild/ (opens in a new tab)
- The Register: Attackers pummel critical WordPress vuln to create all sorts of mischief. https://www.theregister.com/security/2026/07/20/attackers-pummel-critical-wordpress-vuln-to-create-all-sorts-of-mischief/5275265 (opens in a new tab)
- NVD (NIST): CVE-2026-63030. https://nvd.nist.gov/vuln/detail/CVE-2026-63030 (opens in a new tab)
- Sucuri: Vulnerability & Patch Roundup: July 2026. https://blog.sucuri.net/2026/07/vulnerability-patch-roundup-july-2026.html (opens in a new tab)
- Shield Security: ShieldNotes 117. https://getshieldsecurity.com/blog/shieldnotes-117/ (opens in a new tab)
- Cybersecurity News: Critical WordPress Plugin Vulnerability: miniOrange. https://cybersecuritynews.com/wordpress-plugin-vulnerability-miniorange/ (opens in a new tab)
- NVD (NIST): CVE-2026-57807 (miniOrange OAuth SSO). https://nvd.nist.gov/vuln/detail/CVE-2026-57807 (opens in a new tab)
- GBHackers: Critical WordPress Plugin Backdoor Exposes 20,000 Sites. https://gbhackers.com/critical-wordpress-plugin-backdoor-exposes-20000-sites/ (opens in a new tab)
- WordPress.org: WordPress 7.0.3 Release. https://wordpress.org/news/2026/08/wordpress-7-0-3-release/ (opens in a new tab)
- NVD (NIST): CVE-2026-64638. https://nvd.nist.gov/vuln/detail/CVE-2026-64638 (opens in a new tab)
- WordPress.org: WordPress 7.0.4 Release. https://wordpress.org/news/2026/08/wordpress-7-0-4-release/ (opens in a new tab)
The facts in this article have been cross-checked against official primary sources (WordPress.org release notes, the NVD CVE database and the advisory from the National Cyber Security Centre Finland at Traficom). The Advanced Responsive Video Embedder case has no official CVE identifier, so it is based on reporting by Wordfence and GBHackers. AI was used in writing this article.
Article contents
- July's big case: “WP2Shell” broke straight into WordPress core
- July's plugin wave: more than 100 new vulnerabilities
- A separate and still unresolved risk: miniOrange's SSO plugin
- When the update itself is the malware
- August: WordPress 7.0.3 and 7.0.4 closed more holes
- Why and how did the attacks happen?
- What can we learn from this?
- Frequently asked questions
- Sources:
Read more
All articlesBlog
I built a website for five grand, but it would cost you 30 grand
We are one of Finland's leading WordPress agencies. Yet our new website does not contain a single line of WordPress. Most of the site was built by me, the CEO, and I can't code. In this article I explain what that proves and what it doesn't.
19 Nov 2026 · 8 min read
Blog
The EU AI Act gets real: what does Article 50 mean in practice?
The EU AI Act has been talked about for a long time. Now it finally means something. And it means something for all of us who build or use AI-based digital services. On 2 August 2026, the Act's transparency obligations in Article 50 started to apply. In practice, they affect every organisation that uses AI in its website, online service or content.
9 Jun 2026 · 3 min read
Blog
How does a website move business forward?
We say we build websites that move our clients' business forward. What does that actually mean? In practice, we first agree on the goals together with the client, such as growing the number of leads. Only then do we design and build what serves those goals. Finally, we follow the numbers and keep improving the site based on the results.
13 Feb 2026 · 2 min read
