WordPress: Five Critical Flaws Including WPMU DEV Dashboard, Avada, and GiveWP (CVSS 9.8 to 10)
> TL;DR: Wordfence and Patchstack detailed five critical flaws in the WordPress ecosystem on August 29, 2026: an authentication bypass in WPMU DEV Dashboard when Hub SSO is enabled (CVE-2026-76581, CVSS 9.8), arbitrary file write in the Avada theme via Fusion Builder (CVE-2026-18431, 9.8), exposure of raw admin password-reset URLs in TranslatePress (CVE-2026-19632, 9.8), unauthenticated escalation to administrator in Pods (CVE-2026-19598, 9.8), and remote code execution in GiveWP through PHP object injection (CVE-2026-82222, CVSS 10). All of them lead, to varying degrees, to complete site takeover.
The five flaws and their activation conditions
Each flaw has precise prerequisites — knowing them determines whether your site is exposed:
GiveWP: PHP deserialization ending in RCE
The GiveWP case deserves attention because its mechanism is a classic that returns every year in the PHP ecosystem. Patchstack's analysis describes a three-ingredient chain: a "safe unserialize" helper that does not actually strip serialized objects, a donation flow feeding that helper attacker-controlled data, and a gadget chain in code GiveWP ships in production. When the three align, PHP object injection becomes command execution.
The root causes are generic and reach far beyond this plugin: trusting a serialization sanitizer that does not clean objects, unserializing data read back from the database as if it were trusted, and shipping development-only libraries into production where they provide ready-made gadget chains.
Why this matters to you
WordPress sites at SaaS companies are rarely the product itself — they are the marketing site, the blog, the careers pages, and documentation spaces. But they often share the main domain, integrations (application forms, double opt-in, CRM webhooks), and sometimes shared accounts. A compromised WordPress admin on the root domain is a credible pivot toward the product: script injection on pages your users browse, form submission capture, transactional email spoofing. For an HR team running a recruiting site on WordPress with GiveWP or TranslatePress, this list should trigger an immediate version check.
What to do
The broader lesson: "specific conditions" protect less than you think
Four of these five flaws have activation conditions (SSO enabled, secondary locale, published form). Across a fleet of sites, those conditions always match a few real instances — precisely the ones using the feature most. Protection-by-configuration-rarity is not a strategy: it is a delay of a few weeks between your site and the first scanner that tries the right combination. Fast updates of active plugins, even on "secondary" sites, remain one of the best security ROIs on the web.
Building HR, payroll, or recruiting software? CleanIssue performs security audits for HR SaaS in real-world conditions, no source code access needed. For a first read of your exposure, start with an external review of your application.
Related articles
Three adjacent analyses to keep exploring the same attack surface.
WordPress CVE-2026-8206: unauthenticated admin takeover via the Kirki plugin
The Kirki WordPress plugin (6.0.0 to 6.0.6) lets an unauthenticated attacker take over an administrator account via a broken password reset. CVSS 9.8, active exploitation, ~150,000 sites exposed. Fix: 6.0.7.
Adobe Campaign Classic CVE-2026-48449: a perfect 10.0 RCE in enterprise marketing
A maximum-severity authorization flaw (CVSS 10.0) in Adobe Campaign Classic allows arbitrary code execution without user interaction. Paired with CVE-2026-48448 (SQL injection, 8.6) allowing arbitrary file reads. Fixed in ACC v7 build 9398. Not yet exploited.
Next.js CVE-2026-75604: Unauthenticated RCE on Windows Plus an AVIF Heap Overflow — Emergency Patches from August 25
Vercel fixed two critical Next.js flaws: a path traversal exploitable without authentication when the server runs on a Windows filesystem (CVE-2026-75604, CVSS 9.0, no workaround), and a heap buffer overflow in the sharp/libheif chain triggered by a malicious AVIF image (CVSS v4 9.5). Fixes: 15.5.24 and 16.3.3. Vercel-hosted deployments are protected.
Sources
Related services
If this topic maps to a real risk in your stack, these are the most relevant CleanIssue audits.