If you’ve been building WordPress plugins and using AI tools in your workflow, there’s some genuinely exciting news. On 20th March 2026, WordPress.org launched an official MCP (Model Context Protocol) server for the Plugin Directory — meaning you can now submit, validate, and manage your plugins directly from AI tools like Claude, Cursor, or VS Code with AI capabilities, without ever touching the web form.
This post covers what it is, what it can do, and walks you through getting set up and submitting your first plugin using it.
What Is the WordPress Plugin Directory MCP Server?
MCP (Model Context Protocol) is a standard that lets AI assistants connect to external services as “tools” — think of it like giving your AI an API it can use on your behalf. The WordPress.org MCP server connects AI tools directly to the Plugin Directory, giving them the ability to validate, check, and submit plugins.
It was announced by Konstantin Obenland on the Make WordPress Meta blog and is built on top of WordPress Core’s Abilities API and the MCP Adapter plugin.
Crucially, this doesn’t bypass any review process. All plugins submitted through the MCP server go through exactly the same manual review as those submitted via the web form. The plugin guidelines apply regardless of how the code was produced — you’re still responsible for everything your AI generates.
What Can It Do?
The MCP server exposes three main tools to your AI assistant:
- Validate Readme — Checks your
readme.txtfor errors, warnings, and suggestions before you submit. Great for catching formatting issues early. - Get Plugin Status — Retrieves the current review status of a submitted plugin, including any feedback from the review team.
- Submit Plugin — Submits a new plugin for review, or updates a submission that’s still in the review queue.
Beyond those tools, your AI assistant also gets access to reference documents — the Plugin Guidelines, Developer FAQ, and reserved slugs list — so it can give you accurate, up-to-date guidance as you work.
There are also guided prompts built in for structured workflows: preparing a plugin for submission, running Plugin Check locally, and addressing reviewer feedback.
What You’ll Need
- A WordPress.org account
- An MCP-compatible AI client (Claude Desktop, Claude Code, Cursor, Codex, or VS Code with AI)
- Node.js 18 or later installed
Getting Set Up
Setup is a single command. Open your terminal and run:
npx -y @wporg/mcpThis will:
- Open your browser to authorise the connection with your WordPress.org account
- Create an application password for secure access
- Automatically detect and configure your installed MCP clients
If your client isn’t auto-detected, you can follow the manual setup guide in the Plugin Developer Handbook.
Once it’s done, your AI assistant will have access to the Plugin Directory tools. You can verify this in Claude Desktop or Cursor by checking the available tools — you should see the WordPress.org plugin tools listed.
Walkthrough: Building and Submitting an Example Plugin
Let’s walk through building a simple plugin from scratch and using the MCP server to submit it. We’ll build a plugin called Simple Read Time — it adds an estimated reading time to the top of all your posts automatically.
Step 1: Create Your Plugin File
Create a folder called simple-read-time and inside it create your main plugin file, simple-read-time.php:
<?php
/**
* Plugin Name: Simple Read Time
* Plugin URI: https://example.com/simple-read-time
* Description: Displays an estimated reading time at the top of each post.
* Version: 1.0.0
* Author: Your Name
* Author URI: https://example.com
* License: GPLv2 or later
* License URI: https://www.gnu.org/licenses/gpl-2.0.html
* Text Domain: simple-read-time
*/
if ( ! defined( 'ABSPATH' ) ) {
exit;
}
function srt_estimated_reading_time( $content ) {
if ( ! is_single() ) {
return $content;
}
$word_count = str_word_count( strip_tags( $content ) );
$minutes = max( 1, (int) ceil( $word_count / 200 ) );
$label = sprintf(
_n( 'Estimated reading time: %d minute', 'Estimated reading time: %d minutes', $minutes, 'simple-read-time' ),
$minutes
);
$reading_time = '<p class="srt-reading-time"><em>' . esc_html( $label ) . '</em></p>';
return $reading_time . $content;
}
add_filter( 'the_content', 'srt_estimated_reading_time' );Step 2: Create Your readme.txt
Every plugin needs a properly formatted readme.txt. This is what generates your plugin’s page on WordPress.org. Create it in the same folder:
=== Simple Read Time ===
Contributors: yourwordpressusername
Tags: reading time, posts, content
Requires at least: 6.0
Tested up to: 6.7
Stable tag: 1.0.0
Requires PHP: 7.4
License: GPLv2 or later
License URI: https://www.gnu.org/licenses/gpl-2.0.html
Displays an estimated reading time at the top of each single post.
== Description ==
Simple Read Time automatically calculates and displays an estimated reading time
at the top of every single post on your WordPress site. It assumes an average
reading speed of 200 words per minute.
No settings page, no bloat — just drop it in and it works.
== Installation ==
1. Upload the plugin files to `/wp-content/plugins/simple-read-time/`
2. Activate the plugin through the 'Plugins' screen in WordPress
3. Reading time will appear automatically on all single posts
== Frequently Asked Questions ==
= Can I change the reading speed? =
Not in this version, but it's on the roadmap for 1.1.0.
= Where does the reading time appear? =
At the top of each single post, above the content.
== Changelog ==
= 1.0.0 =
* Initial releaseStep 3: Validate Your Readme with the MCP Server
Before submitting, ask your AI assistant to validate your readme. In Claude Desktop or Cursor, you can simply say:
“Validate my WordPress plugin readme — here’s the content: [paste your readme.txt]”
The MCP server’s Validate Readme tool will check for:
- Missing required fields (like
Stable tagorLicense) - Formatting errors in section headers
- Common warnings like missing screenshots or FAQ sections
Fix any issues it flags before moving on.
Step 4: Run Plugin Check Locally
The Plugin Check plugin is the official tool for verifying your plugin meets WordPress.org directory requirements. Install it on a local WordPress instance, activate it, and run it against your plugin before submitting. Your AI assistant can walk you through this step by step — just ask it to guide you through running Plugin Check on your plugin.
Since September 2024, Plugin Check has been integrated into the WordPress.org review process directly, and it’s been shown to reduce issues at approval time by 41%. Running it locally before you submit is well worth doing.
Step 5: Zip Your Plugin
Zip the plugin folder (not just the files inside it — the folder itself should be inside the zip):
zip -r simple-read-time.zip simple-read-time/Step 6: Submit via the MCP Server
Now for the new part. In your AI client, ask it to submit your plugin:
“Submit my WordPress plugin for review. The plugin is called Simple Read Time.”
Your AI assistant will use the Submit Plugin tool to handle the submission. It will ask for any details it needs — your plugin zip, readme content, etc. — and submit it to the WordPress.org review queue on your behalf.
Step 7: Check Your Submission Status
After submitting, you can check in on the review at any time by asking your AI assistant:
“What’s the current status of my Simple Read Time plugin submission?”
The Get Plugin Status tool will retrieve the current review state and any feedback from the plugin review team. Review typically takes between 1 and 10 business days.
Step 8: Respond to Feedback (if needed)
If the review team has questions or asks for changes, your AI assistant can help you address them. It has access to the full Plugin Guidelines and Developer FAQ, so you can ask it to explain any feedback and suggest how to fix it. Once you’ve made changes, use Submit Plugin again to update your submission.
After Approval: SVN Still Applies
One thing to note: the MCP server handles the submission and review workflow, but once your plugin is approved, you’ll still need to use Subversion (SVN) to push your actual plugin files to the WordPress.org repository. The Plugin Directory is a hosting platform, not just a listing — your code needs to live in the SVN repo for the plugin to be downloadable.
WordPress.org will email you with your SVN repository URL once approved. From there, check out the repo, add your files to /trunk, commit, and your plugin will go live.
What’s Next for the MCP Server?
The architecture is designed to grow. WordPress.org has signalled that more tools are planned, potentially including role-specific tools for plugin reviewers and forum moderators. Community members can also contribute new tools directly as pull requests to the WordPress.org GitHub repository.
If you run into any issues with the current tools, open a ticket on Meta Trac.