WordPress Sites Face Active Exploitation Risks During Core Security Updates, Researchers Warn

Millions of WordPress websites around the globe could still face exposure to remote code execution attacks during the precise moment they install crucial security updates, according to new research. The vulnerability, which has been under active exploitation in the wild, highlights a critical design flaw in how the world’s most popular content management system handles core package deployment and temporary file extraction.

The security flaw, tracked officially under the identifier CVE-2026-119977, impacts a massive span of WordPress installations ranging from version 4.7 up to version 7.1.3. The issue stems from the mechanism WordPress utilizes when performing a core update using a complete package. During this routine maintenance process, the platform automatically unpacks a copy of the fresh version into a temporary folder established within the site’s internal wp-content/upgrade directory.

Crucially, this temporary directory includes a functional copy of the WordPress installer, and standard web server configurations usually allow external web access to it. According to Umut Bayram, the Picus Security researcher who originally uncovered the critical flaw, an unauthorized attacker with no account privileges whatsoever can potentially determine or predict the exact name of this temporary folder. If the malicious actor strikes at the exact right moment before WordPress deletes the temporary copy, they can successfully execute arbitrary code directly on the underlying server.

Recognizing the severity of the threat and observing active exploitation in the wild, the U.S. Cybersecurity and Infrastructure Security Agency (CISA) added the vulnerability to its official catalog of known exploited vulnerabilities. The inclusion places federal and critical infrastructure entities on high alert, prompting immediate calls for heightened monitoring and defensive configurations across web hosting environments.

In response to the discovery, WordPress released a formal security update via version 7.1.4, alongside matching patch releases for older supported branches dating back to version 4.7. However, Bayram’s technical analysis points out a paradoxical security challenge: the very update process used to install the fix can momentarily expose the site to the vulnerability it aims to resolve.

The core of the problem lies in the legacy codebase utilized during the installation phase. When a site initiates an update, the platform unpacks the incoming files using the old, unpatched code currently residing on the server. On installations running WordPress 5.9 or later, this legacy update routine names the temporary directory directly after the downloaded archive file, making the destination name predictable for a brief window of time.

Before the implementation of broader security controls, malicious actors could exploit this predictable window by deploying a crafted configuration file, known as wp-config.php, straight into the temporary copy. When the update procedure finalized, the script would inadvertently overwrite the live site’s authentic configuration file with the malicious payload. While the primary patch successfully prevents this specific configuration-overwrite vector by leveraging updated code to copy files into place, attackers have pivoted their tactics.

Installing the Fix for an Exploited WordPress Flaw Can Expose Some Sites One Last Time

An attacker who manages to intercept the fleeting window of the update process can still configure the temporary copy of WordPress with a bespoke database, log into it, and execute arbitrary PHP code with the elevated permissions of the web server before the system purges the directory. From that unauthorized foothold, malicious actors gain the capability to read, modify, or inject files across the entire live WordPress environment, giving them near-total control over the compromised server architecture.

Who Is Exposed During the Update

Understanding the precise exposure window requires examining how WordPress handles different types of updates. Security analysts emphasize that not every platform update exposes the installer files to external networks. Complete update packages inherently include the full suite of installer files, whereas smaller partial packages—designed specifically for incremental jumps between minor versions—only expose a site if those specific files happen to be modified in the patch.

For websites configured to handle updates automatically in the background, WordPress initiates the remediation via the standard internal updater mechanism, following the identical set of steps as a manual deployment. Security monitoring conducted shortly after the disclosure revealed that routine point updates within a specific branch, such as transitioning from version 7.1.3 to 7.1.4, typically deliver a partial package. Conversely, major jumps from an older software branch, such as moving from version 7.0.7 up to 7.1.4, force the system to rely exclusively on complete packages, widening the potential exposure window.

For legacy installations running versions 4.4 through 5.8, the older updater logic appends randomized characters to the temporary folder’s name. This design effectively obscures the directory path unless the web application inadvertently leaks the location through verbose error messages or misconfigured debugging logs. Once a site successfully reaches version 7.1.4 or an equivalent branch patch, subsequent updates adopt randomized folder naming conventions permanently, neutralizing predictable path targeting.

What to Do Before Updating

To mitigate the inherent risks posed during the update cycle, security researchers have outlined specific defensive strategies for server administrators. The primary and most immediate recommendation is to block external web access to the wp-content/upgrade directory at the web server configuration level, such as through Nginx or Apache rule adjustments. Because WordPress operates directly on the underlying disk storage when managing upgrades, restricting external HTTP access to this specific folder does not interfere with the functionality of core updates.

An alternative, highly secure methodology involves performing manual updates entirely outside of the designated web root—the primary directory served publicly by the web server. System administrators can accomplish this by creating a complete backup of both the site files and the underlying database, unpacking the new release archive in a private directory inaccessible to the web, copying the core files into place while strictly preserving the existing wp-config.php and wp-content directories, and finally executing the database upgrade routine. This out-of-band approach ensures that no temporary installer files are ever instantiated within a web-accessible directory structure.

Installing the Fix for an Exploited WordPress Flaw Can Expose Some Sites One Last Time

Security experts caution that standard command-line utilities, such as the wp core update command provided by WP-CLI, do not offer a bypass for this risk. The utility ultimately invokes the same underlying updater logic and unpacks packages into the identical upgrade directory. Furthermore, legacy installations running WordPress 4.6 and earlier receive no official security updates from the core development team, leaving them entirely unpatched. Researchers stress that installations running these obsolete versions must migrate to a supported software branch to maintain basic security hygiene.

Attacks and What to Check

Although CISA confirmed active exploitation of the flaw in enterprise environments, initial advisories lacked granular details regarding the specific tactics employed by threat actors. However, commercial security telemetry has shed light on the speed and scale of automated exploitation campaigns targeting the ecosystem.

According to data compiled by Wordfence, automated attack probes began a mere nine minutes after the security company deployed a virtual firewall rule designed to intercept exploits on the day of the release. Telemetry identified prominent traffic sources continuously scanning targeted groups of websites for roughly fourteen and a half hours, a duration that security analysts correlate with automated attempts to catch the exact second a vulnerable site initiates its core update routine.

The severity of CVE-2026-119977 is rated at 8.1 out of 10 on the Common Vulnerability Scoring System (CVSS) scale, reflecting its potential for complete remote code execution, though individual lab tests indicate that the vulnerable update window can close within mere seconds. Concurrently, functional proof-of-concept exploit code has begun circulating in public code repositories, with some script authors claiming compatibility with predictable temporary folder names utilized during complete-package updates.

For web administrators managing sites that remained on vulnerable versions during the height of the exploitation campaigns, security firms recommend conducting thorough forensic audits. Key indicators of compromise include unexpected modifications to core files, unauthorized database user accounts, unfamiliar PHP scripts introduced into upload directories, and anomalous entries within web server access logs corresponding with the timestamps of past core updates.

Subsequent clarifications from security researchers emphasize that the ultimate determining factor for exposure is not whether an update package is partial or complete, but rather whether the specific updater codebase already resident on the host site has been thoroughly patched against the underlying directory-prediction logic.