Critical Avada WordPress Theme Flaw (CVE-2026-18431) Enables Zero-Click RCE

Critical Avada WordPress Theme Flaw (CVE-2026-18431) Enables (TL-2026-2157) is a critical-severity software vulnerability scored CVSS 9.8, first published 2026-08-26 and last reviewed 2026-08-29. It has no confirmed attribution, affects ThemeFusion Avada (WordPress theme), references 1 CVE (CVE-2026-18431), maps to 13 MITRE ATT&CK techniques (T1027, T1059.004, T1068), and is covered by 9 detection rules and 15 indicators of compromise.

Key facts for TL-2026-2157

Threat ID
TL-2026-2157
Severity
CRITICAL
CVSS
9.8 (CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H)
Status
PATCHED
Category
VULNERABILITY
First published
2026-08-26
Last reviewed
2026-08-29
Attribution confidence
LOW
Motivation
UNKNOWN
Detection rules
9
Indicators of compromise
15
Updates
2026-08-29 · revalidated 1× · latest source

Malware and tooling in Critical Avada WordPress Theme Flaw (CVE-2026-18431) Enables

Malware and tooling: Argus, PRISM, Argus, PRISM

A CVSS 9.8 unauthenticated remote code execution flaw in the Avada WordPress theme (<=7.16) and bundled Fusion Builder plugin (<=3.16) lets an attacker chain six authorization and file-handling weaknesses in the built-in 'Fusion Patcher' update subsystem into an arbitrary PHP file write. Wordfence's Argus AI research agent found, reproduced, and privately proved the chain in about two hours; ThemeFusion shipped patched 7.16.1/3.16.1 builds on 2026-08-25 with no public PoC and no reported in-the-wild exploitation.

How Critical Avada WordPress Theme Flaw (CVE-2026-18431) Enables works

CVE-2026-18431 (CWE-862, Missing Authorization) affects the Avada WordPress theme up to and including version 7.16 together with its bundled Fusion Builder page-builder plugin up to and including version 3.16, when both are installed and active. The root cause sits in the Fusion Patcher, Avada's built-in mechanism for fetching update instructions from outside the site and applying them to local files and options -- functionally a self-update channel, and correspondingly powerful when abused. Wordfence describes an unauthenticated six-step exploit chain: (1) a public-facing request exposes attacker-controlled input; (2) that input reaches internal behavior not intended for anonymous users; (3) a privileged component (the Fusion Patcher) is invoked outside its intended context; (4) the request data is used to influence trusted server-side state for the duration of the request; (5) an administrative/maintenance operation is reached with insufficient authorization checks; and (6) file-handling controls fail to adequately constrain the location and content of files the Patcher writes, allowing an arbitrary PHP file to be created on disk. Because every link in the chain is required, breaking any single one blocks the demonstrated path -- but no individual step looks obviously malicious in isolation, which is why Wordfence characterizes this class of flaw as evading naive signature-based scanning and single-request WAF rules.

Independent detection-engineering analysis (SecurityArsenal) mapping the disclosed step pattern onto known WordPress exploitation primitives assesses the likely underlying mechanics as: an unauthenticated AJAX (`admin-ajax.php`) or REST (`/wp-json/`) request that leaks state, writes attacker-controlled data into a transient/option, or bypasses a nonce check (steps 1-2); a second request that weaponizes that state -- e.g. registering a malicious shortcode, injecting a serialized PHP object (PHP object injection), or setting an arbitrary option (steps 3-4); and a final trigger that coerces the application into including or evaluating attacker-controlled PHP via template injection, an `include`/`require` on a now-writable path, or a file write into `wp-content/uploads/` (steps 5-6). This is a plausible reconstruction consistent with the publicly confirmed chain shape, not Wordfence's withheld technical detail.

The vulnerability was discovered by Argus, a new single-target agentic AI research tool Wordfence built specifically to find long, multi-step exploit chains that would take a skilled human researcher weeks or months to connect (if they connected all six links at all). Argus found and reproduced the chain, and wrote a working proof-of-concept, unattended in approximately two hours. It is the second such AI research agent Wordfence has publicly disclosed in 2026, alongside PRISM (launched April 2026), which has logged over 300 vulnerabilities and caught a WordPress.org-hosted plugin supply-chain backdoor in under two hours in July 2026. Wordfence has deliberately withheld the technical specifics of how Argus works and has not published exploit code, arguing the same agentic techniques that help defenders find and fix flaws quickly would help attackers exploit them just as quickly.

Disclosure timeline: the CVE was reserved and Argus found/reproduced the chain on 2026-07-30, the same day Wordfence deployed a firewall rule to Premium/Care/Response customers; ThemeFusion was notified on 2026-08-05 and acknowledged the report on 2026-08-10; patched Avada 7.16.1 and Fusion Builder 3.16.1 shipped on 2026-08-25 (Avada's changelog and blog post credit Wordfence by name, though the vendor's public post withholds function-level technical detail and points users to the full changelog/help docs instead); the free-tier Wordfence firewall rule is scheduled for 2026-08-29 under Wordfence's standard 30-day premium-to-free delay. CVE-2026-18431 was published in NVD on 2026-08-26 with CVSS 3.1 base score 9.8 (AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H). As of publication there is no public proof-of-concept, no confirmed in-the-wild exploitation, and the CVE does not appear in the CISA Known Exploited Vulnerabilities catalog. Given the >1 million-site install base of Avada (narrowed in practice to sites running both Avada and Fusion Builder active together), the critical unauthenticated pre-auth nature of the flaw, and the demonstrated speed at which agentic tooling can now both find and (in principle) weaponize such chains, defenders are urged to patch immediately rather than wait for evidence of active exploitation.

Because full technical detail is withheld, defensive guidance published by third-party detection engineers focuses on the likely post-exploitation artifacts and follow-on behavior of an arbitrary-PHP-write primitive rather than the exact request chain: monitoring POST traffic to `admin-ajax.php` and the WordPress REST API for webshell payload markers (including legacy `preg_replace()` `/e`-modifier code-execution patterns); watching for new `.php`/`.phtml`/`.phar`/`.php7`/`.php8` files written into `wp-content/uploads/`; auditing `wp-content/mu-plugins/` for planted must-use-plugin backdoors (a common WordPress persistence mechanism once arbitrary file write is achieved); checking the `wp_users` table for unauthorized new administrator accounts; checking `wp-config.php` for signs of database-credential harvesting/exposure; alerting on PHP-FPM/Apache/Nginx worker processes spawning shell child processes; and hunting for outbound connections from web-server worker processes to non-RFC1918 (external) addresses as a beaconing/C2 indicator. Site owners are advised to update both Avada and Fusion Builder immediately, verify version numbers (`wp-content/themes/Avada/style.css`) across every affected site, and treat the Fusion Patcher -- and self-update/patch-delivery subsystems generally -- as sensitive attack surface going forward, since a vulnerability in a self-patching mechanism grants an attacker whatever that mechanism can do.

MITRE ATT&CK techniques used in TL-2026-2157

Defense Evasion

T1027 Obfuscated Files or Information; T1140 Deobfuscate/Decode Files or Information; T1211 Exploitation for Defense Evasion

Execution

T1059.004 Unix Shell

Privilege Escalation

T1068 Exploitation for Privilege Escalation

Command and Control

T1071.001 Web Protocols; T1105 Ingress Tool Transfer

Persistence

T1136.001 Local Account; T1505.003 Web Shell

Initial Access

T1190 Exploit Public-Facing Application

Collection

T1213 Data from Information Repositories

Credential Access

T1552.001 Credentials In Files

Resource Development

T1584.004 Compromise Infrastructure

Affected products and versions in Critical Avada WordPress Theme Flaw (CVE-2026-18431) Enables

  • ThemeFusion — Avada (WordPress theme)
    Vulnerable versions: <= 7.16
    Fixed in: 7.16.1
  • ThemeFusion — Fusion Builder (WordPress plugin, bundled with Avada)
    Vulnerable versions: <= 3.16
    Fixed in: 3.16.1

Remediation for Critical Avada WordPress Theme Flaw (CVE-2026-18431) Enables

Patches

  • Avada 7.16.1 (released 2026-08-25) — fixes the Fusion Patcher remote code execution vulnerability
  • Fusion Builder 3.16.1 (released 2026-08-25)

Immediate actions

  • Update the Avada theme to 7.16.1 or later
  • Update the Fusion Builder plugin to 3.16.1 or later
  • Verify the deployed version in wp-content/themes/Avada/style.css and the Fusion Builder plugin listing after updating, across every WordPress site running Avada

Workarounds

  • Deactivate Fusion Builder / Avada Builder if immediate patching is not possible
  • Wordfence Premium/Care/Response firewall rule available since 2026-07-30; free-tier Wordfence firewall rule scheduled for 2026-08-29

Longer-term hardening

  • Block direct PHP execution inside wp-content/uploads/ via .htaccess/Nginx rules to reduce webshell impact if a similar file-write flaw recurs
  • Deny PHP execution in wp-includes/ where feasible
  • Monitor POST traffic to admin-ajax.php and the WordPress REST API for webshell payload markers (eval(, base64_decode(, system(, passthru(, shell_exec(, assert(, preg_replace()
  • Alert on PHP-FPM/Apache/Nginx worker processes spawning shell child processes (/bin/sh, /bin/bash, curl, wget, nc, python)
  • Hunt for outbound connections from web-server worker processes to non-RFC1918 addresses as a C2-beaconing indicator
  • Audit wp-content/mu-plugins/ for planted must-use-plugin backdoors
  • Audit the wp_users table for unauthorized administrator accounts created after the disclosure window
  • Check wp-config.php for signs of database-credential exposure/harvesting
  • Treat self-update/patch-delivery subsystems (like Fusion Patcher) as sensitive attack surface and reduce patch SLA for critical CMS RCEs given AI-accelerated vulnerability research timelines

CVEs associated with Critical Avada WordPress Theme Flaw (CVE-2026-18431) Enables

CVE-2026-18431

Weaknesses (CWE) in Critical Avada WordPress Theme Flaw (CVE-2026-18431) Enables

CWE-862

Timeline of Critical Avada WordPress Theme Flaw (CVE-2026-18431) Enables

  • Wordfence deploys a firewall rule covering the vulnerability to Premium, Care, and Response customers.
  • Wordfence's Argus AI research agent finds and reproduces the six-step exploit chain and writes a working internal proof-of-concept, unattended, in approximately two hours.
  • CVE-2026-18431 is reserved (per NVD/CVE record metadata), coinciding with Argus's discovery of the chain.
  • Wordfence notifies ThemeFusion of the vulnerability.
  • ThemeFusion acknowledges receipt of the vulnerability report.
  • ThemeFusion releases Avada 7.16.1 and Fusion Builder 3.16.1, fixing the Fusion Patcher remote code execution vulnerability and crediting Wordfence in the changelog and blog post.
  • BleepingComputer and other outlets publish coverage of the vulnerability; Wordfence continues to withhold full technical/exploit detail.
  • CVE-2026-18431 is published in the NVD with CVSS 3.1 score 9.8 (Critical).
  • NVD's CVE-2026-18431 record last-modified timestamp updates.
  • WPScan's vulnerability database entry last-updated timestamp updates; CVE-2026-18431 is confirmed still absent from the CISA Known Exploited Vulnerabilities catalog as of this date.
  • Wordfence's free-tier firewall rule is scheduled to release, following the standard 30-day premium-to-free protection delay.

Update history for TL-2026-2157

Sources cited for Critical Avada WordPress Theme Flaw (CVE-2026-18431) Enables

More in vulnerability

Detection coverage for TL-2026-2157

As of 2026-08-29, Threadlinqs Intelligence publishes 9 detection rule(s) for TL-2026-2157 across Splunk SPL, Microsoft KQL and Sigma, covering 15 indicator(s) of compromise. The whole corpus is readable without an account; a free account unlocks full detection query text in Splunk SPL, Microsoft KQL and Sigma; paid tiers add raw indicator values, correlation and the MCP server. Threadlinqs MCP server · View plans.

Threadlinqs Intelligence — Real-Time Threat Detection Platform

[ 0 threats ] [ 0 det ] [ CRIT: 0 ] [ HIGH: 0 ]
// threat_feed
$ sort --newest
Showing all threats

Latest Threats