Nonces Csrf Protection logo

Nonces Csrf Protection

Organization
wpultimatesecurity
nonces-csrf-protection

Use when handling any form submission, AJAX request, admin-post action, settings page, link that triggers an action, or any other user-initiated request in a WordPress plugin or theme. Generates nonces with wp_nonce_field / wp_create_nonce and verifies them with check_admin_referer, check_ajax_referer, or wp_verify_nonce, always paired with a capability check, to prevent CSRF. Apply proactively whenever code accepts or acts on a request.

Overview

Publisherwpultimatesecurity
RepositoryWordPress-Security-Skills
Skill namenonces-csrf-protection
Stars
54
Forks
3
Bundled files
2
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 2 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by wpultimatesecurity on GitHub. Read the source before you install it.

Installation

Install the Nonces Csrf Protection AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/wpultimatesecurity/WordPress-Security-Skills/tree/dev/skills/nonces-csrf-protection
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/wpultimatesecurity/WordPress-Security-Skills.git /tmp/WordPress-Security-Skills
mkdir -p .claude/skills
cp -r /tmp/WordPress-Security-Skills/skills/nonces-csrf-protection .claude/skills/nonces-csrf-protection
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Nonces Csrf Protection in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Nonces Csrf Protection on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Nonces Csrf Protection is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Nonces & CSRF protection in WordPress

When to use this skill

Use this skill whenever code accepts or acts on a request that changes state or performs a privileged action. Concretely:

  • HTML <form> submissions (admin or front-end) handled via admin-post.php, admin_post_*, or a settings page.
  • admin-ajax.php handlers (wp_ajax_*, wp_ajax_nopriv_*) and fetch()/jQuery AJAX.
  • Action links such as "Delete", "Approve", "Activate" that carry a query argument.
  • Settings API pages (the Settings API adds a nonce automatically — verify you are not bypassing it).
  • Any custom endpoint reached over $_GET / $_POST / $_REQUEST.

If a request only reads public data and changes nothing, a nonce is not required — but the moment it writes, deletes, or triggers a side effect, it is mandatory.

A nonce is not authentication and not authorization. It proves the request came from a page your site generated, defeating CSRF. You still need a capability check (current_user_can) to prove the user is allowed to do the thing. Always use both. See the capability-permission-checks skill.

Core principles (and why they matter)

  1. Every state-changing request needs a nonce. CSRF works by tricking a logged-in admin's browser into submitting a forged request. A nonce the attacker cannot guess breaks that attack.
  2. Nonce + capability are a pair, never a substitute. The nonce says "this request came from us"; the capability check says "this user may do this." A nonce alone lets any logged-in subscriber perform admin actions; a capability check alone leaves you open to CSRF.
  3. Tie the nonce to a specific action. Use a unique, descriptive action string (e.g. delete_widget_42) rather than a generic one. A nonce scoped to "delete widget 42" cannot be replayed to delete widget 99 or to do something unrelated.
  4. Verify on the server, every time. Generating a nonce does nothing on its own. The security comes from verifying it on the receiving side before acting.
  5. Fail closed. If verification fails, stop — wp_die(), return a 403, or send a JSON error. Never fall through to the action.
  6. Nonces expire (default ~24h) and are per-user. They are single-action tokens, not long-lived secrets. Don't store or reuse them; regenerate per page render.

Step-by-step implementation

  1. Generate the nonce where the request originates:
    • In a form: wp_nonce_field( 'my_action', 'my_nonce' ) (prints a hidden field).
    • In a URL: wp_nonce_url( $url, 'my_action', 'my_nonce' ).
    • For JS/AJAX: wp_create_nonce( 'my_action' ), passed to the script via wp_localize_script() or wp_add_inline_script().
  2. Send it with the request (hidden field, query arg, or AJAX payload / header).
  3. Verify it on the server first, before reading other input or acting:
    • Form via admin-post.php: check_admin_referer( 'my_action', 'my_nonce' ).
    • AJAX: check_ajax_referer( 'my_action', 'nonce' ).
    • Manual / REST-ish: wp_verify_nonce( $nonce, 'my_action' ) and branch on the result.
  4. Check capability immediately after: if ( ! current_user_can( 'manage_options' ) ).
  5. Then wp_unslash() + sanitize the input, do the work, and escape any output.
  6. Fail closed on any failure with wp_die() or wp_send_json_error().

Supporting references

ReferenceLoad when
Nonce / CSRF verification checklistBefore final verification of the nonce / csrf verification controls.
Secure AJAX nonce flowImplementing the privileged AJAX flow from nonce generation through verification, authorization, and JSON response.

Common AI mistakes / anti-patterns

Mistake 1 — Verifying the nonce but skipping the capability check

php
// ❌ Insecure: nonce proves the request shape, NOT that the user is allowed.
add_action( 'admin_post_delete_thing', function () {
    check_admin_referer( 'delete_thing' );
    delete_thing( absint( $_POST['id'] ) ); // any logged-in user can reach this
} );
php
// ✅ Secure: nonce AND capability.
add_action( 'admin_post_delete_thing', function () {
    check_admin_referer( 'delete_thing' );
    if ( ! current_user_can( 'manage_options' ) ) {
        wp_die( esc_html__( 'You are not allowed to do this.', 'my-plugin' ), 403 );
    }
    delete_thing( absint( wp_unslash( $_POST['id'] ?? 0 ) ) );
} );

Mistake 2 — Reading wp_verify_nonce() as a boolean and ignoring the result

wp_verify_nonce() returns 1, 2, or false — not a clean boolean, and crucially the call has no effect unless you branch on it.

php
// ❌ Insecure: result is computed and thrown away; the action always runs.
wp_verify_nonce( $_POST['my_nonce'], 'my_action' );
save_settings();
php
// ✅ Secure: branch on the result and fail closed.
$nonce = isset( $_POST['my_nonce'] ) ? sanitize_text_field( wp_unslash( $_POST['my_nonce'] ) ) : '';
if ( ! wp_verify_nonce( $nonce, 'my_action' ) ) {
    wp_die( esc_html__( 'Security check failed.', 'my-plugin' ), 403 );
}
save_settings();

Mistake 3 — Generic / reused action strings

php
// ❌ Insecure: a single global action lets a valid nonce be replayed across endpoints.
wp_nonce_field( 'nonce' );           // action = "nonce"
check_admin_referer( 'nonce' );
php
// ✅ Secure: specific, scoped action string (include the object id when relevant).
wp_nonce_field( 'delete_widget_' . $widget_id, 'widget_nonce' );
check_admin_referer( 'delete_widget_' . $widget_id, 'widget_nonce' );

Mistake 4 — wp_ajax_nopriv_* for a privileged action

wp_ajax_nopriv_* fires for logged-out visitors. Wiring a sensitive action there exposes it to the public.

php
// ❌ Insecure: settings save reachable by anonymous users.
add_action( 'wp_ajax_nopriv_save_api_key', 'save_api_key' );
php
// ✅ Secure: privileged actions use wp_ajax_* only, plus nonce + capability inside.
add_action( 'wp_ajax_save_api_key', 'save_api_key' );
function save_api_key() {
    check_ajax_referer( 'save_api_key', 'nonce' );
    if ( ! current_user_can( 'manage_options' ) ) {
        wp_send_json_error( array( 'message' => 'Forbidden' ), 403 );
    }
    // ...sanitize + save...
    wp_send_json_success();
}

Mistake 5 — Echoing the nonce without escaping, or building it by hand

php
// ❌ Insecure / fragile: manual markup, unescaped output.
echo '<input type="hidden" name="n" value="' . wp_create_nonce( 'act' ) . '">';
php
// ✅ Secure: let core print the (already-escaped) field, or escape explicitly.
wp_nonce_field( 'act', 'n' );
// If you must build a URL by hand, escape it:
$url = wp_nonce_url( admin_url( 'admin-post.php?action=act' ), 'act', 'n' );
echo '<a href="' . esc_url( $url ) . '">' . esc_html__( 'Do it', 'my-plugin' ) . '</a>';

Mistake 6 — Trusting check_ajax_referer() to also authorize

check_admin_referer() and check_ajax_referer() only verify the nonce (and referer). They do not check capabilities. Add current_user_can() yourself.

Correct code examples

Admin form submitted to admin-post.php

php
<?php
/**
 * Render the settings form.
 */
function my_plugin_render_form() {
    if ( ! current_user_can( 'manage_options' ) ) {
        return;
    }
    ?>
    <form method="post" action="<?php echo esc_url( admin_url( 'admin-post.php' ) ); ?>">
        <input type="hidden" name="action" value="my_plugin_save">
        <?php wp_nonce_field( 'my_plugin_save', 'my_plugin_nonce' ); ?>
        <input type="text" name="api_key" value="">
        <?php submit_button( __( 'Save', 'my-plugin' ) ); ?>
    </form>
    <?php
}

/**
 * Handle the submission.
 */
add_action( 'admin_post_my_plugin_save', 'my_plugin_handle_save' );
function my_plugin_handle_save() {
    // 1. Verify the nonce (fails closed via wp_die on mismatch).
    check_admin_referer( 'my_plugin_save', 'my_plugin_nonce' );

    // 2. Verify capability.
    if ( ! current_user_can( 'manage_options' ) ) {
        wp_die( esc_html__( 'You are not allowed to do this.', 'my-plugin' ), 403 );
    }

    // 3. Unslash + sanitize input.
    $api_key = isset( $_POST['api_key'] )
        ? sanitize_text_field( wp_unslash( $_POST['api_key'] ) )
        : '';

    // 4. Do the work.
    update_option( 'my_plugin_api_key', $api_key );

    // 5. Redirect back safely.
    wp_safe_redirect( add_query_arg( 'updated', 'true', wp_get_referer() ) );
    exit;
}

For the full AJAX flow (PHP handler + JS), see references/secure-ajax-handler.php.

Header-based and JSON nonces for REST / modern fetch

When a form is not available, send the nonce in a header or JSON body. The REST API expects X-WP-Nonce with a nonce created via wp_create_nonce( 'wp_rest' ).

js
// JavaScript: fetch with X-WP-Nonce header.
fetch( '/wp-json/my-plugin/v1/thing', {
    method: 'POST',
    headers: {
        'Content-Type': 'application/json',
        'X-WP-Nonce': myPluginData.nonce, // created with wp_create_nonce( 'wp_rest' )
    },
    body: JSON.stringify( { id: 1 } ),
} );
js
// Or use @wordpress/api-fetch, which attaches the REST nonce automatically.
import apiFetch from '@wordpress/api-fetch';

apiFetch( {
    path: 'my-plugin/v1/thing',
    method: 'POST',
    data: { id: 1 },
} );

On the server, the REST API verifies the X-WP-Nonce header when you use a real permission_callback. Do not use __return_true for state-changing REST routes.

Related: see the ajax-security skill for the admin-ajax flow and the rest-api-security skill for REST-specific permission handling.

Checklist

Before finishing any request-handling code, confirm:

  • Every state-changing request generates a nonce (wp_nonce_field, wp_nonce_url, or wp_create_nonce).
  • The action string is specific and includes the target object id where relevant.
  • The server verifies the nonce before doing anything else (check_admin_referer / check_ajax_referer / branched wp_verify_nonce).
  • A capability check (current_user_can) runs alongside the nonce check.
  • wp_verify_nonce() results are branched on; the action does not run on failure.
  • Privileged AJAX uses wp_ajax_* (not wp_ajax_nopriv_*).
  • On failure the code fails closed (wp_die, 403, or wp_send_json_error).
  • Input is wp_unslash()-ed and sanitized after verification.
  • Nonce values printed to the page are escaped (or produced by core helpers).

Official references

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Nonces Csrf Protection AI skill do?

Use when handling any form submission, AJAX request, admin-post action, settings page, link that triggers an action, or any other user-initiated request in a WordPress plugin or theme. Generates nonces with wp_nonce_field / wp_create_nonce and verifies them with check_admin_referer, check_ajax_referer, or wp_verify_nonce, always paired with a capability check, to prevent CSRF. Apply proactively whenever code accepts or acts on a request.

Why use Nonces Csrf Protection on TypingMind?

Because you install it once and use it with any model. Nonces Csrf Protection is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Nonces Csrf Protection in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/wpultimatesecurity/WordPress-Security-Skills/tree/dev/skills/nonces-csrf-protection. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Nonces Csrf Protection?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Nonces Csrf Protection?

As many as you like. As long as a model supports skills, you can use Nonces Csrf Protection with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Nonces Csrf Protection AI skill free?

Yes. It is published on GitHub by wpultimatesecurity under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇