Skip to content

Light version: pages open quickly, without the 3D graphics. You are seeing the light version because the 3D graphics would be slow on this device. The content is the same.

Turn on 3D

Preparing the graphics. This can take a moment on the first visit.

Henry Lõiv

Web design & development

Website Security 16 min read

Rank Math PRO malware and Node.js security fixes: what to check

A combined guide to the Rank Math PRO supply-chain incident and October Node.js security fixes: affected versions, eleven CVEs, practical checks and recovery.

Check two different parts of your website’s software stack. Some Rank Math PRO downloads contained malware; Node.js has released fixes for ten vulnerabilities. A site that executed an affected plugin package needs incident recovery, while a vulnerable Node runtime needs a verified update and an exposure review.

Sources checked on 22 October 2026. This guide combines the two issues because a website can depend on both. It does not establish that they share an attacker, or that your website has been compromised.

The practical question is which code ran, where it ran and what it could access. Your WordPress dashboard, public website, development laptop, build service and background workers can have different software versions and different credentials. A reassuring result from one environment says little about the others.

This article covers CVE-2026-111862 for Rank Math PRO and all ten CVEs in Node.js’s 21 October release. It separates confirmed malicious software from vulnerability disclosures, then explains how to turn a security announcement into an accountable maintenance task.

Start with the component you actually use

An ordinary WordPress website runs PHP on its server. Node.js may run an application beside it, or only build the theme’s JavaScript and CSS before deployment. The presence of a package.json file does not tell you which arrangement you have.

Your setupWhat needs checkingWho normally has the answer
WordPress with Rank Math PROThe installed package, download history and whether an affected copy was activatedWebsite maintainer and plugin account owner
WordPress with the free Rank Math pluginThe free plugin is outside this PRO malware incident; identify other components separatelyWebsite maintainer
WordPress or a static site with a Node buildDeveloper machines, CI runners and container images used during buildsDeveloper or deployment maintainer
A production Node applicationRunning web processes, APIs, workers, scheduled tasks and deployment imagesApplication operator or hosting provider
Several client websites on shared infrastructureEach installation and the access shared between themHosting administrator and agency maintainer

Assign a named person to each row that applies. “The hosting company handles security” is too vague if the host manages the server but your agency manages premium plugins and a separate service builds the website.

For example, a shop can have WordPress on managed PHP hosting, a theme built on a developer’s laptop and a Node worker that sends order information to another system. Updating WordPress addresses none of those other components automatically. Equally, the fact that Node is installed on a laptop does not establish that the public WordPress site exposes a Node server.

Rank Math PRO: identify the affected package, not just today’s version

The official CVE record limits affected Rank Math PRO 3.0.123 packages to downloads or updates from 15 October 2026, 20:02 UTC to 16 October, 07:12 UTC. CVE-2026-111862 record. In Estonia, that was 15 October at 23:02 to 16 October at 10:12, while UTC+3 was in force. The free WordPress.org plugin is excluded. Vendor notice. The clean replacement is 3.0.124. Official PRO release history.

Version 3.0.123 originally appeared on 7 October. That matters: the version label alone does not identify a tainted download. The same label can appear in an installation obtained before the incident window. Use the package’s origin and acquisition history when deciding whether the incident applies. Official PRO release history.

CVE-2026-111862 describes embedded malicious code in the affected PRO packages. Wordfence assigns a critical CVSS 3.1 score of 9.8. This is a supply-chain compromise: the concern is malicious code delivered through a legitimate plugin distribution route, rather than a normal plugin feature that an administrator can make safe by changing one setting. Wordfence vulnerability record.

Build a short installation history. Record the domain, plugin edition, current version, acquisition date if known, activation history and source of that information. Useful evidence includes deployment records, update-management logs, the original archive, vendor correspondence and backups that show when files changed. File timestamps are clues; copying or restoring files can change them.

A current clean version does not answer whether an earlier affected package ran. Activation can also happen after the download window: an archive downloaded during the window can be installed later. An empty memory of the update is not a reliable negative result; ask the account owner or maintainer to resolve the history.

Choose the right Rank Math response

The vendor distinguishes three cases. A downloaded archive that was never installed does not, by itself, establish website compromise. For an installed but never activated affected copy, delete the affected plugin files, install a clean replacement and scan. An activated affected copy requires treating the site as compromised and following recovery guidance. Vendor response branches.

Preserve the facts and contain the affected installation

Before a responder begins removing files, record the discovery time, relevant domains, known package history and actions already taken. Preserve available web-server, hosting, deployment and authentication records. Keep investigation copies access-controlled; database exports and configuration files can contain customer information and credentials. WordPress’s response guidance emphasizes documenting symptoms and recent changes so an investigation starts from evidence. WordPress incident-response guidance.

Containment is a technical decision for the person responsible for the affected environment. If serving the site allows malicious code to keep executing, the responder may need to restrict access at the hosting or application layer while preserving evidence. A maintenance page implemented inside an already compromised application is not an independent security boundary.

For a shop, identify how legitimate orders received after a candidate backup will be preserved. For a membership site, identify recent registrations and account changes. Restoring a database without planning that reconciliation can replace a security incident with missing business records. Do not merge entire compromised tables back into a clean restoration simply to recover recent transactions; choose and review the necessary business records.

Restore or clean the whole affected environment

The target is an installation whose files, database and access can be trusted again. Replacing the visible plugin folder does not establish that outcome. Persistent code, altered configuration or an additional account can survive elsewhere. Patchstack’s assessment also explains why a web-application firewall rule cannot remove malicious software that has already arrived in a plugin package. Patchstack assessment.

Have the responder choose between a sufficiently early clean backup and a controlled rebuild or cleanup. A backup must be assessed against the installation timeline; “the latest backup” may already contain the affected code. Files and database belong in the recovery plan together. WordPress’s hardening guidance treats backups, restricted access and monitoring as complementary controls, rather than a single product that certifies a site as safe. WordPress hardening guidance.

Persistent backdoors can be recreated after incomplete removal. Wordfence’s incident classification. Have the responsible specialist follow the current vendor recovery procedure rather than a detached deletion checklist.

Replace exposed access and check its use

Make a credential inventory based on what the affected web process could reach. Think beyond the WordPress administrator password: database access, WordPress security keys and salts, mail delivery, payment integrations, cloud APIs, deployment access and credentials shared with another installation may sit within that boundary.

For each item, record its owner, where the replacement must be installed, how the old credential will be revoked and how normal service will be checked. Rotate from a trusted environment and avoid delivering new secrets into a system where persistence remains unresolved. Review authentication and integration records for misuse; changing a password does not explain what happened before the change.

These are response-planning steps, not a claim that every listed service lost data. The impact depends on the permissions and information actually available to the compromised process. WordPress’s security guidance supports limiting database, filesystem and account privileges so that one installation has less access to other resources. Access-control guidance.

What WordPress checksums can and cannot establish

In a controlled investigation environment, a developer can compare core files with the official release:

wp core verify-checksums

This is a comparison of WordPress core files, not a complete review of the database, uploads, users or server. Use the correct release and locale where required. The command deliberately runs before WordPress loads. Core checksum command documentation.

Plugin verification is a separate check:

wp plugin verify-checksums --all

It uses WordPress.org’s plugin checksums. A premium plugin may have no matching public checksum set; an unavailable result is not proof of either malware or cleanliness. A legitimate custom change can also cause a mismatch that needs interpretation. Compare premium packages with a trusted official distribution through the responder’s workflow. Plugin checksum command documentation.

Do not treat a successful command, one scanner report or a normally loading homepage as a certificate that the entire environment is clean. Preserve the outputs alongside the recovery record and explain which parts of the system they actually covered. Commands that load WordPress components should be assessed before use on a suspected compromised installation. Official explanation of checksum limits.

Node.js: install a corrected runtime in the branch you use

Node.js published security releases on 21 October 2026. The first fixed versions for the maintained branches in that release are:

Node.js branchFirst fixed release in this security updateWhat to deploy
22.x22.23.4This release or a later appropriate supported security release
24.x24.22.1This release or a later appropriate supported security release
26.x26.12.1This release or a later appropriate supported security release

The project’s highest severity classification for this release is high. It does not describe every listed issue as a critical remote takeover. Check the 22.23.4 release, 24.22.1 release and 26.12.1 release for the official builds.

A distribution or commercial supplier can backport a security correction while retaining an older-looking upstream version. In that case, obtain the supplier’s advisory for the relevant CVEs and the complete patched package revision; do not accept the short version string as the only evidence. Red Hat’s explanation of security backports.

For an older, unsupported branch, the absence of a row is not assurance. Arrange migration to a supported release or obtain a specific supported-supplier remediation statement. The normal choice for a production application should follow Node’s supported LTS lifecycle and application compatibility, rather than a hurried jump to whichever major version has the largest number. Node.js release lifecycle.

All ten Node.js CVEs in the October release

The table groups the faults by the part of an application worth reviewing. Branches refer to the affected maintained release lines before their corrected releases. These are exposure questions, not ten declarations that an application has been breached. The official Node.js security notice lists the complete release scope.

CVENode severity and branchesWhat needs attention
CVE-2026-111497High; 22, 24, 26Brotli decoding can write outside its intended memory region and permit code execution. Review untrusted compressed inputs.
CVE-2026-105482High; 22, 24, 26DNS resolution truncates names containing a NUL character. Review hostname validation and resolution boundaries.
CVE-2026-105483High; 24, 26TLS hostname checks can be bypassed with fullwidth-digit IP representations. Review outbound connection identity.
CVE-2026-106069High; 22, 24, 26TLS hostname checking mishandles host values interpreted loosely as IP addresses.
CVE-2026-105481Medium; 22, 24, 26A keep-alive socket can process a response written before its request. Review HTTP client response handling.
CVE-2026-106072Medium; 22, 24, 26Non-canonical IPv4 names can confuse certificate validation between DNS identities and IP endpoints.
CVE-2026-106074Medium; 22, 24, 26Startup snapshots and single-executable applications can preserve randomness caches and repeat generated values.
CVE-2026-106070Low; 26ZIP handling can retain invalid header state and return attacker-controlled prefix data on a subsequent read.
CVE-2026-106071Low; 26Wildcard matching can grant too much filesystem access under the Permission Model.
CVE-2026-106073Low; 22, 24, 26fs.realpathSync can bypass the Permission Model’s intended filesystem restriction.

Brotli: consider incoming requests and outgoing responses

A Node application or one of its libraries may decode compressed data received from a user, an upstream service or a fetched resource. The relevant question is whether attacker-controlled Brotli data reaches the affected decoder. Merely showing that the application serves JavaScript does not answer that question. Node.js compression API documentation.

Vercel’s bulletin discusses exploitation through compressed data and explains the difference between protecting incoming traffic and processing data fetched by the application. A front-door firewall cannot inspect every response the server fetches elsewhere. Review upload processors, proxies and fetch paths alongside public request handlers. Code execution would occur with the privileges of the Node process; those privileges and reachable secrets determine the possible consequences. Vercel’s Node.js security bulletin.

DNS and TLS: inspect the identity you meant to contact

The hostname issues concern the relationship between a supplied host, the address it resolves to and the certificate that authenticates the connection. They deserve attention in applications that accept a user-supplied URL or retrieve data from changing destinations.

Write down the intended trust boundary. If an import tool is meant to contact only an approved partner, which code decides that the destination is approved, and which runtime performs the actual connection? A validation check and a connection that interpret a string differently are an integration problem worth reviewing. Apply the runtime fixes rather than trying to invent a blacklist of unusual hostname spellings.

Do not disable certificate verification to make an integration work after an update. That would remove the identity check the application needs. Review connection errors, redirects and proxy behavior with the responsible developer and keep the expected destination explicit. Node.js TLS identity-check documentation.

Randomness snapshots are application artifacts

The snapshot issue concerns startup state and single-executable applications. It is not a claim that restoring a virtual-machine backup makes all random values insecure. If your team produces these artifacts, identify the build inputs and recreate the affected artifacts as part of deploying corrected code.

An ordinary application that never produces or consumes those artifacts has a different exposure question. Keep the distinction in the maintenance record, even when the decision is still to install the security release for the other fixes.

Permission restrictions do not replace operating-system isolation

Two fixes concern Node’s Permission Model. Review them where an application relies on Node’s own filesystem restrictions. Such restrictions are one layer in the design; service-account privileges, filesystem ownership and container boundaries still deserve separate attention. A corrected Node binary should be deployed before relying on the corrected behavior. Node.js Permission Model documentation.

Update the executable that actually runs the work

The following commands are inventory examples. They report the Node selected in the environment where they run; they do not scan an application for exploitation:

node --version
node -p 'JSON.stringify({version: process.version, executable: process.execPath})'

The first shows the runtime version. The second also identifies that process’s executable path. Run the check in each relevant environment, through its normal service or deployment workflow where practical. A terminal opened under your account may select a different binary from a process manager, scheduled job or container. Avoid publishing the resulting local paths if they reveal information you do not intend to share. Node process version and executable documentation.

Create an inventory with one row per environment: production web service, worker, scheduled task, CI job, build image and developer workstation. Record the current runtime, its supplier, its owner and the corrected target. One row called “the server” will miss an older worker image running on the same host.

Replace the runtime using the installation method your environment already manages. Then restart or redeploy the processes that execute it and verify the result through those processes. Installing a new binary on disk does not replace the code already loaded by a running service. Keep the deployment and rollback method understandable to the person operating it.

For containers, refresh the relevant base image and build a new application image. Recreate or redeploy the running containers from that image; editing a Dockerfile alone changes no existing container. Retain the image identifier or digest alongside the running-version check so the deployment can be traced. Docker image-rebuild guidance.

An npm update or audit is a different task

Project dependencies and the Node executable have separate update paths. npm audit examines dependency information for known package vulnerabilities. Its output should not be used as proof that the Node runtime underneath the project contains these security fixes. A clean dependency report and an old runtime can coexist. npm audit documentation.

Keep dependency changes deliberate during an urgent runtime update. Record whether you changed the runtime, the lockfile, the base image or all three. If application behavior changes afterwards, that separation makes diagnosis and rollback more precise.

Development and CI machines also need review. A static production site may have no Node service at all, while its build machine can still process downloaded data and hold deployment credentials. Assess what the build actually executes and what credentials it can read. That is a reason to maintain the build environment, not evidence that its published HTML already contains malware.

Verify business functions and the security outcome separately

After an authorized update or recovery, check the real paths people use: administrator sign-in, contact-form delivery, authentication callbacks, file handling, checkout and external integrations as applicable. A successful homepage response does not exercise an order worker or a scheduled synchronization job.

Record functional results separately from incident findings. “Checkout completed in the test environment” answers an availability question. “An affected plugin package was never activated, supported by deployment records” answers an exposure question. “The rebuilt worker reports the corrected runtime” answers a deployment question. None substitutes for the others.

If suspicious accounts, files, tasks or outbound activity are found, preserve the relevant records and extend the investigation. Patching a vulnerability closes the documented flaw; it does not reverse actions performed before the patch. Avoid turning a version inventory into a claim that no data was accessed.

A message to send your maintainer

The following is a suggested request for evidence, not a statement that the recipient’s systems are compromised:

Please review our website and its build/deployment environments for the October 2026 Rank Math PRO incident and Node.js security release. Confirm which Rank Math edition and package were installed, when the package was obtained and whether an affected copy was activated. For each Node environment, identify the running version and supplier, the corrected target and the deployment completed. Where exposure is confirmed or uncertain, explain the investigation, containment, recovery and credential-review plan. Please retain the evidence and distinguish completed work from open questions.

Ask for the result by environment, not a single reply saying “everything is updated.” For shared hosting or an agency-managed portfolio, also clarify whether the answer covers neighboring installations, staging sites and reusable build images.

Keep ownership explicit after the urgent work. Someone should own premium-plugin notices, someone should own runtime lifecycle changes, and someone should be able to restore a tested backup. Those can be the same person, but the responsibilities should not disappear between hosting, development and marketing accounts.

The critical PHP fixes and eMudhra certificate incident affect other layers of the stack. A Node release may also update bundled trust material, but that does not establish that a separate PHP installation, application trust store or server certificate has been remediated.

If you need help identifying what your existing website runs and who maintains it, describe the system and its symptoms through the website maintenance enquiry. Include the relevant domains and service names; send passwords, keys and customer records only through an agreed secure process.

Written by Henry Lõiv

Keep reading

All articles⁠
  1. Best AI agents in 2026: Grok Bot vs Dots and EU availability Compare AI agents by the work they can handle and whether you can use them in Estonia or the EU. Artificial Intelligence 5 min read
  2. AI Agent, Chatbot or Automation: What Does Your Business Need? A chatbot is an interface; an agent is one way to handle the work behind it. Follow a fictional customer enquiry to choose a practical system for your business. Artificial Intelligence 8 min read
  3. Claude Code vs Codex for Frontend Development: A Practical Comparison A documentation-based comparison with a runnable frontend exercise and a blank scorecard. No measured winner is claimed. Artificial Intelligence 9 min read

Hi!

I really love the sea, yachts, sunsets and starry skies.

Through this little 3D world, I can share a small part of myself with you.

Thank you for opening this bottle and reading my letter.

Henry

Let’s talk about your project

Tell me about your project

Tell me what you’d like to build or improve. You don’t need a finished brief — we’ll work out the scope and next steps together.

01 How can I help?
02 How many pages do you need?
03 Which extras do you need? Choose all that apply
04 When does it need to be ready?
05 Project details
06 Your contact details

Enquiry received

Thank you!

Your enquiry has been received. I usually reply by email within 2 working days.