Ajax Security logo

Ajax Security

Organization
wpultimatesecurity
ajax-security

Use when registering or handling WordPress AJAX over admin-ajax.php - wp_ajax_{action} / wp_ajax_nopriv_{action} hooks, JavaScript that posts to admin_url('admin-ajax.php'), or wp.apiFetch / fetch calls to custom actions. Verifies the nonce with check_ajax_referer, gates the action with current_user_can, unslashes and sanitizes every field, and replies with wp_send_json_success / wp_send_json_error. Prevents CSRF, broken access control, and injection on the AJAX surface.

Overview

Publisherwpultimatesecurity
RepositoryWordPress-Security-Skills
Skill nameajax-security
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 Ajax Security 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/ajax-security
  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/ajax-security .claude/skills/ajax-security
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Ajax Security 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 Ajax Security 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 Ajax Security 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.

AJAX security

When to use this skill

Use this skill whenever code registers or handles an AJAX action through WordPress's admin-ajax.php endpoint:

  • add_action( 'wp_ajax_{action}', ... ) for logged-in users.
  • add_action( 'wp_ajax_nopriv_{action}', ... ) for logged-out visitors.
  • JavaScript posting to admin_url( 'admin-ajax.php' ) with jQuery, fetch, or wp.apiFetch.
  • Returning JSON, HTML fragments, or any server-computed data from an AJAX handler.

If the request changes state or exposes privileged data, it needs a nonce and a capability check. wp_ajax_nopriv_* is intentionally unauthenticated — never use it for admin actions.

Related: see the nonces-csrf-protection skill for the full nonce lifecycle and the rest-api-security skill for REST endpoints.

Core principles (and why they matter)

  1. wp_ajax_* vs wp_ajax_nopriv_* are different trust boundaries. nopriv fires for anonymous visitors; use it only for truly public actions. Privileged actions must use wp_ajax_*.
  2. Nonce + capability for privileged actions. check_ajax_referer() checks a CSRF token, not authenticated origin or authorization; current_user_can() checks permission.
  3. Sanitize every input field. $_POST values in AJAX are just as attacker-controllable as any other request. wp_unslash() then sanitize to type before use.
  4. Exit through wp_send_json_* or wp_die(). Never echo raw output and fall through. JSON responses keep the contract explicit and prevent accidental HTML/PHP leakage.
  5. Scope the nonce to the action. A nonce for my_plugin_delete_item_42 cannot be replayed to delete item 99.
  6. Fail closed. Any failed check returns a JSON error or dies; the handler never continues to act.
  7. Throttle anonymous and expensive actions before the work. Public nonces can be obtained and replayed by attackers; they are not rate limits. Transient counters are best-effort load shedding only: read/increment/write is non-atomic and cached entries can disappear early. Strict limits need an atomic shared backend or an edge/server limiter with defined windows, trusted client identity, and failure policy.

Step-by-step implementation

  1. Register the handler with the correct hook (wp_ajax_* for authenticated, wp_ajax_nopriv_* for public).
  2. Enqueue the script and pass the AJAX URL + nonce via wp_localize_script() or wp_add_inline_script().
  3. In the handler, verify the nonce first with check_ajax_referer( $action, $query_arg, false ). Passing false lets you return a clean JSON error instead of dying with -1.
  4. Check current_user_can() immediately after the nonce.
  5. wp_unslash() and sanitize every $_POST / $_GET field, using the right sanitizer for the type (absint, sanitize_text_field, sanitize_key, sanitize_email, etc.).
  6. Enforce the abuse policy before queries, mail, or other expensive actions; return wp_send_json_error( ..., 429 ) when rejected. Use an atomic limiter where bypass would create a security or cost risk; a transient is only a best-effort supplement.
  7. Perform the authorized action.
  8. Return wp_send_json_success() or wp_send_json_error() and stop.

Supporting references

ReferenceLoad when
AJAX security checklistBefore final verification of AJAX security controls.
Secure AJAX handlerImplementing the privileged admin-ajax.php profile-update flow with nonce, capability, input validation, and JSON responses.

Common AI mistakes / anti-patterns

Mistake 1 — No nonce verification

php
// ❌ Insecure: any site can POST to this action and trigger the handler.
add_action( 'wp_ajax_my_plugin_vote', function () {
    $post_id = absint( $_POST['post_id'] );
    update_post_meta( $post_id, '_votes', get_post_meta( $post_id, '_votes', true ) + 1 );
    wp_send_json_success();
} );
php
// ✅ Secure: verify nonce, then capability, then act.
add_action( 'wp_ajax_my_plugin_vote', 'my_plugin_ajax_vote' );
function my_plugin_ajax_vote() {
    if ( ! check_ajax_referer( 'my_plugin_vote', 'nonce', false ) ) {
        wp_send_json_error( array( 'message' => __( 'Security check failed.', 'my-plugin' ) ), 403 );
    }
    if ( ! current_user_can( 'edit_posts' ) ) {
        wp_send_json_error( array( 'message' => __( 'Forbidden.', 'my-plugin' ) ), 403 );
    }
    $post_id = isset( $_POST['post_id'] ) ? absint( wp_unslash( $_POST['post_id'] ) ) : 0;
    if ( $post_id <= 0 ) {
        wp_send_json_error( array( 'message' => __( 'Invalid post.', 'my-plugin' ) ), 400 );
    }
    update_post_meta( $post_id, '_votes', get_post_meta( $post_id, '_votes', true ) + 1 );
    wp_send_json_success();
}

Mistake 2 — Using wp_ajax_nopriv_* for a privileged action

php
// ❌ Insecure: anonymous users can now save settings.
add_action( 'wp_ajax_nopriv_my_plugin_save_settings', 'my_plugin_save_settings' );
php
// ✅ Secure: privileged actions use wp_ajax_* only.
add_action( 'wp_ajax_my_plugin_save_settings', 'my_plugin_save_settings' );
function my_plugin_save_settings() {
    check_ajax_referer( 'my_plugin_save_settings', 'nonce' );
    if ( ! current_user_can( 'manage_options' ) ) {
        wp_send_json_error( array( 'message' => __( 'Forbidden.', 'my-plugin' ) ), 403 );
    }
    // ...sanitize + save...
    wp_send_json_success();
}

Mistake 3 — Echoing raw $_POST or building HTML without escaping

php
// ❌ Insecure: reflected XSS.
echo '<div>' . $_POST['message'] . '</div>';
php
// ✅ Secure: escape at the point of output, even inside a JSON payload.
$message = sanitize_text_field( wp_unslash( $_POST['message'] ?? '' ) );
wp_send_json_success( array( 'html' => '<div>' . esc_html( $message ) . '</div>' ) );

Mistake 4 — Forgetting wp_unslash()

php
// ❌ Risky: WordPress slashes superglobals; data may contain escaped quotes.
$label = sanitize_text_field( $_POST['label'] );
php
// ✅ Secure: unslash first, then sanitize.
$label = isset( $_POST['label'] ) ? sanitize_text_field( wp_unslash( $_POST['label'] ) ) : '';

Mistake 5 — Trusting is_user_logged_in() instead of a capability

php
// ❌ Insecure: every logged-in user, including subscribers, can run this.
if ( is_user_logged_in() ) {
    delete_option( 'my_plugin_config' );
}
php
// ✅ Secure: check a real capability.
if ( ! current_user_can( 'manage_options' ) ) {
    wp_send_json_error( array( 'message' => __( 'Forbidden.', 'my-plugin' ) ), 403 );
}

Mistake 6 — Unthrottled public action

php
// ❌ Insecure: anonymous endpoint, no rate limit — unbounded cost per visitor.
add_action( 'wp_ajax_nopriv_my_plugin_search', 'my_plugin_ajax_search' );
function my_plugin_ajax_search() {
    $q = isset( $_POST['q'] ) ? sanitize_text_field( wp_unslash( $_POST['q'] ) ) : '';
    wp_send_json_success( my_plugin_run_expensive_query( $q ) );
}
php
// Best-effort only: public, read-only search with bounded output.
add_action( 'wp_ajax_nopriv_my_plugin_search', 'my_plugin_ajax_search' );
function my_plugin_ajax_search() {
    if ( ! isset( $_POST['q'] ) || ! is_string( $_POST['q'] ) || strlen( $_POST['q'] ) > 200 ) {
        wp_send_json_error( array( 'message' => 'Invalid query.' ), 400 );
    }
    $q = sanitize_text_field( wp_unslash( $_POST['q'] ) );
    if ( '' === $q ) {
        wp_send_json_error( array( 'message' => 'Query required.' ), 400 );
    }
    // Fixed minute buckets; counters can race or be evicted. Not a strict quota.
    $ip     = $_SERVER['REMOTE_ADDR'] ?? '';
    $window = intdiv( time(), MINUTE_IN_SECONDS );
    $key    = 'my_plugin_rl_' . wp_hash( 'search|' . $ip . '|' . $window );
    $count  = (int) get_transient( $key );
    if ( $count >= 20 ) {
        wp_send_json_error( array( 'message' => __( 'Too many requests.', 'my-plugin' ) ), 429 );
    }
    set_transient( $key, $count + 1, MINUTE_IN_SECONDS );

    // The counter is checked before the query, never after sending the response.
    $ids = get_posts( array(
        's'           => $q,
        'post_type'   => 'post',
        'post_status' => 'publish',
        'numberposts' => 10,
        'fields'      => 'ids',
    ) );
    wp_send_json_success( $ids );
}

This read-only public search does not use a nonce as an abuse control. State-changing or privileged handlers still need their CSRF and capability checks. Fixed windows allow bursts across boundaries; parallel requests can lose counter increments, and cache eviction/storage failure can reset this best-effort counter. Do not use it as the sole control for mail sending, paid APIs, voting integrity, or brute-force defense.

REMOTE_ADDR identifies the direct peer: behind a proxy it may identify the proxy, not the visitor, and shared NATs group users together. Only trust forwarded addresses after a configured trusted proxy strips client-supplied headers and supplies a validated address; never read arbitrary X-Forwarded-For directly. Strict distributed limits need an atomic shared operation (including expiry, e.g. a Redis script), or an edge/server limiter on all relevant routes. See authentication-session-security for login-specific considerations.

Correct code examples

A complete, copy-paste-ready AJAX handler (PHP + JavaScript) is in references/secure-ajax-handler.php.

Checklist

  • The action is registered on wp_ajax_* for privileged flows or wp_ajax_nopriv_* only when the action is truly public.
  • The handler verifies the nonce with check_ajax_referer() before acting.
  • A current_user_can() check runs immediately after nonce verification.
  • Every $_POST / $_GET field is wp_unslash()-ed and sanitized to type.
  • The handler exits via wp_send_json_success(), wp_send_json_error(), or wp_die().
  • Output inside JSON responses is escaped for its context (esc_html, esc_attr, esc_url).
  • The nonce action string is specific and scoped (e.g., includes an object id).
  • Secrets are never returned to the browser or logged.
  • Abuse controls run before expensive actions; strict limits use atomic shared or edge enforcement, not transients.
  • Client identity accounts for trusted proxies/shared IPs; arbitrary forwarded headers are not trusted.

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 Ajax Security AI skill do?

Use when registering or handling WordPress AJAX over admin-ajax.php - wp_ajax_{action} / wp_ajax_nopriv_{action} hooks, JavaScript that posts to admin_url('admin-ajax.php'), or wp.apiFetch / fetch calls to custom actions. Verifies the nonce with check_ajax_referer, gates the action with current_user_can, unslashes and sanitizes every field, and replies with wp_send_json_success / wp_send_json_error. Prevents CSRF, broken access control, and injection on the AJAX surface.

Why use Ajax Security on TypingMind?

Because you install it once and use it with any model. Ajax Security 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 Ajax Security in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/wpultimatesecurity/WordPress-Security-Skills/tree/dev/skills/ajax-security. 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 Ajax Security?

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 Ajax Security?

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

Is the Ajax Security 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 👇