CVE-2026-130121: AI Puffer WordPress plugin lets AI comment replies carry stored XSS

The AI Puffer WordPress plugin is vulnerable to stored cross-site scripting through AI-generated comment replies in all versions up to 2.5.28, allowing unauthenticated attackers to inject malicious scripts when the Comment Auto-Reply automation is enabled.

Christopher Ort
By Christopher Ort · 2 min read
Share
A computer screen displaying a WordPress dashboard with comment moderation panel, showing multiple user comments awaiting review
Image: AI-generated, i10X
On this page
  1. Quick Take
  2. In plain English
  3. What happened
  4. What you should do
  5. What it means
  6. Sources

A stored cross-site scripting vulnerability in the AI Puffer WordPress plugin allows unauthenticated attackers to inject malicious scripts through AI-generated comment replies. The flaw affects all versions up to and including 2.5.28.

i10X Quick Take

Quick Take

  1. What: Stored XSS vulnerability in AI Puffer WordPress plugin via AI-generated comment replies
  2. Affected: All versions up to and including 2.5.28, only when Comment Auto-Reply automation is configured
  3. Severity: High (CVSS 7.2)
  4. Status: Fix committed in WordPress plugin changeset 3807151; versions after 2.5.28 carry it
  5. Action: Update the plugin and review Comment Auto-Reply automation settings

In plain English

A WordPress plugin called AI Puffer, which uses AI to automatically reply to comments, has a security flaw. When the plugin’s automatic comment reply feature is turned on, an attacker can post a comment designed to trick the AI into generating a reply that contains hidden malicious code. That code then runs in the browser of anyone who views the comment thread.

What you should do: If you use this plugin and have enabled automatic AI replies to comments, you should update it immediately. If you don’t use this plugin or haven’t turned on the auto-reply feature, you are not affected.

What happened

The AI Puffer plugin for WordPress, which provides MCP server functionality for Claude and ChatGPT along with an AI chatbot, contains a stored cross-site scripting vulnerability in its comment auto-reply feature. The vulnerability exists in the insert_comment_reply_logic() function, which saves AI-generated replies using wp_insert_comment() without properly sanitizing the AI model’s output.

WordPress normally applies kses filtering to comment content to prevent malicious code, but the plugin bypasses this protection when inserting AI-generated replies. An unauthenticated attacker can exploit this by submitting a comment crafted to manipulate the AI model into reproducing an HTML payload in its automated reply. The malicious script then executes whenever any user views the page containing the injected comment.

The vulnerability is only exploitable when an administrator has configured a Comment Auto-Reply automation task with the feature active. The CVSS score is 7.2, classified as high severity.

What you should do

Administrators running the AI Puffer plugin should update to the latest release, since every version up to and including 2.5.28 is affected. The fix was committed in WordPress plugin changeset 3807151. Site owners should also review their Comment Auto-Reply automation settings and consider temporarily disabling the feature until the update is applied. After updating, administrators may want to review existing comments for any suspicious content that may have been injected before the patch.

What it means

This vulnerability highlights a new attack surface introduced by AI-powered automation features in content management systems. When AI models generate content that is directly inserted into web pages without proper sanitization, they can become a vector for traditional web vulnerabilities like XSS. The issue demonstrates that AI-generated content requires the same security controls as user-generated content, and that developers must not assume AI output is inherently safe. The vulnerability’s exploitability depends on specific configuration choices, limiting its impact to sites that have actively enabled the auto-reply feature.

AI plugins for WordPress have had a rough month: we recently covered a privilege escalation in the MCP Server for WordPress plugin and an earlier admin takeover flaw in the same plugin. New cases are tracked in our AI Security hub.

Sources

  • NVD CVE-2026-130121
  • WordPress Plugin Trac: insert_comment_reply_logic() function
  • WordPress Plugin Changeset 3807151
  • Wordfence Threat Intelligence
Christopher Ort

Christopher Ort

Christopher Ort handles SEO, AI agent workflows and PR & partnerships at i10x.ai, the all-in-one AI workspace built by i10X Global Pte. Ltd. in Singapore. He is a graduate of Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU). He is the author of “The Style Penalty”, a study published on SSRN that analyses 1,576 evaluation decisions by four frontier AI models to show how AI-written CVs fare in automated hiring screens; the research has been covered by heise online and t3n.

Editor's picks

  1. CVE-2026-132121: Any website can hijack a local Supergateway and run your MCP tools
  2. CVE-2026-132122: agentic-radar scans can leak your OpenAI API key to a malicious repository
  3. CVE-2026-119220: Dokan MCP endpoint regex flaw lets attackers bypass WooCommerce permissions
  4. Reflection AI Gates Laser, Its Autonomous Penetration Testing Agent, Behind a Trusted Access Program
  5. BrightAI Raises $97 Million to Replace Scheduled Infrastructure Inspections With Physical AI