Introduction to securely developing plugins

With WordPress now powering over 40% of the web, website owners need to think more about online security. Almost every week a new plugin vulnerability is found and patched, leading to concerns about the security of WordPress.

Fortunately for the new plugin developer, the WordPress developer handbook has an extensive section on Plugin Security. If followed, plugin vulnerabilities can be drastically reduced, and the entire ecosystem is protected.

This tutorial will introduce the new developer to the first 5 areas of plugin security, using easy-to-understand examples.

Learning outcomes

1. Securing (sanitizing) input
2. Data validation
3. Securing (escaping) output
4. Preventing untrusted requests
5. Checking User Capabilities

Comprehension questions

1. What are the top three things to keep in mind when thinking about developing plugins securely?
2. What function can be used to secure text input?
3. What function can be used to escape text output?
4. What is the name of the special number used once used to check for untrusted requests.

View Transcript

Hey there, and welcome to Learn WordPress.

In this tutorial, you’re going to learn how to secure your WordPress plugins. You will learn the benefits of ensuring your plugin’s code is secure, what steps to follow to secure your plugins, and where to find more information around plugin security.

Please note that this tutorial was created as an introduction to being security minded when developing plugins. The code examples used in this tutorial are very simplified examples. You should not use any of the code using this tutorial in your plugins. Make sure to read the full documentation on security in the WordPress developer handbook at https://developer.wordpress.org/api/security/ to ensure you follow the correct methods and procedures.

Plugin security is the process of ensuring that your plugin code not only works but also does not introduce any security vulnerabilities. If your plugin has vulnerabilities, it could make any WordPress site that has your plugin installed open to potential attacks, and can lead to the site being compromised. When writing code, it’s important to develop a security mindset and to think about how your code could be used maliciously.

Don’t trust any data, whether it’s user input, third party API data, or even data in your database. Always be checking to make sure it’s valid and safe to use. WordPress has a number of API’s that can help you with common tasks such as sanitising your input, validating data, and escaping output. Rely on using these API’s to help validate and sanitise your data instead of writing your own functions. And finally, keep up to date with common web vulnerabilities and make sure your code is up to date to prevent them.

One of the first steps to take when developing a plugin is to ensure that any user input is sanitised. This means that any data that has been passed to your plugins from the user, such as a form, submission, or URL parameter is checked to make sure it’s safe to use. In this example code, the name and email fields are being posted from a form that the plugin generates, and then save to a custom database table called form submissions. As you can see, the data is being saved directly to the database without any sanitization. This means that if a user were to submit the name of John colon drop table form submissions, the SQL insert query would be run followed by the drop query and the table would be deleted from the database. WordPress has a sanitization API that can be used to sanitise incoming data. You can use the sanitise text field and sanitise email functions to sanitise the name and email fields before they’re using the query

So sanitise textfield

And sanitise email

Read more about the available sanitization functions available to WordPress developers, check out the Sanitising page in the WordPress developer documentation.

Next up is validating data. Validating data is the process of testing it against a predefined pattern or patterns with a definitive result, either valid or invalid. Untrusted data can come from many sources, users, third party API data, even your database data can be considered untrusted, especially if another plugin may have modified it. Even site admins can make a mistake and enter incorrect or insecure data. So it’s important to always be checking. In this example, a deletion function requires that a numeric ID is posted to an admin ajax callback. Here the ID is being used directly in the SQL query without any validation. Again, this means if a user were to submit the ID of one semicolon drop table form submissions, the same SQL drop query would be run after the insert and the table would be dropped. To prevent this you can use PHPs typecasting functionality to ensure that the value of ID is always an integer value. This can be done by adding int before the variable name like so.

Note that this will only work if the first character in the string passed via the post array can be cast an integer, otherwise the value of ID will be zero. In that case, it’s a good idea to update the code to handle this case. So we could say something like if ID is equal to an the same type of zero, then we can return with a JSON response and pass in some kind of error message.

This is only one way to validate integer data. To read more about the various ways to validate your data and your WordPress plugin, check out the section on Validating data in the WordPress developer documentation.

Another aspect of plugin security is to ensure that any information you output to the browser is safe, including any text, HTML, JavaScript code, or data from the database. Even if your plugin is not responsible for the source of the data being displayed, it is responsible for displaying it safely. In this example, this code fetches the form submissions from the database, and then loops through the submission and outputs the submission data into an admin screen in the WordPress dashboard. Here we have three pieces of data that need to be escaped the submission name, and submission email fields, as well as the submission ID. For the name and email fields, we can use the built in WordPress escaping function esc_html.

And we can copy that email.

The ID can be escaped by costing to an integer as you might have done for the data validation.

And then cast to an integer.

Notice that this code follows a key principle of escaping data, and that you escape the data as late as possible. To read more about the various ways to escape data, check out the section on Escaping outputs in the WordPress developer documentation.

Whenever a request is made to functionality that your plugin provides, it’s important to check that the request is valid. This means checking that the request is coming from a trusted source. For example, in your plugin code, you might have a shortcode that renders a form where users can submit their information. The function to render the form might look something like this. When the user submits the form, the data is sent to the server. The data is then processed by the plugin sanitised and stored in the database. Because the form might appear on any page where the shortcode is used, it’s possible that a malicious user could attempt to send a post request to the form either looking for a vulnerability in the plugin or by sending multiple requests of the form. To prevent this, you can check that the request is coming from a trusted source. In other words, the website rendering the form. To do this you can implement something called a nonce or a number used once. First in the form itself you can add a nonce field by using the WP nonce field function and passing a nonce action and nonce name to the function.

Just below the hidden field on the form open up PHP tags and call the WP nonce field function and pass in an action name and then pass in a field name.

When the form is rendered on the front end, a hidden field is added to the form using the nonce name as the ID and Name attributes of the hidden field and the generated nonce as the fields value. This data is posted when the form is submitted. Then in the function that processes the form data, you can verify that the nonce has been sent and is valid by using the WP verify nonce function and passing the post that nonce field and the nonce action to this function. If the result of this verification is false, you can early preventing any further code execution.

So before we sanitise the data, we can say if WP verify nonce and past in WP learn form nonce field posted in the form. And the action name we set up which was WP learn form nonce action. And we specify not to say, if the function returns a false value. In other words, the nonce is not verified, then we can redirect some error page and exit code execution.

Anytime your plugin code makes a web request, be it via redirection to a new URL, posting data to a form or making an Ajax request, you should check that that request is valid using nonces. To read more about how to use nonces in your plugins, check out the section on Nonces in the WordPress developer documentation.

Depending on your plugins functionality, it’s a good idea to restrict certain features only to users with a specific permission level. For example, your plugin may have a function that deletes data from the database. While you might do your best to prevent anyone who doesn’t have permission to run this function, it’s still a good idea to include checks against the situation. WordPress includes a robust user roles and capabilities system, which allows you to either use the default user roles and capabilities or create custom ones. In this case, it could be as easy as only allowing users will have the capability to Manage Site options, which is a standard capability that is included with the administrator role. To do this, you can use the current user can function and check if they have the capability to manage options.

Inside the form submission, nice and early, we can say if not current user can manage options. So in other words, if the user can’t manage options, then we want to return early with some kind of response.

The WordPress developer documentation has a detailed section on user roles and capabilities, which includes a list of all the default capabilities and how to create custom ones. Finally, make sure to read the entry in the WordPress developer documentation on security as it includes all of the examples in this tutorial, as well as additional information on security best practices, common vulnerabilities, and much more example code.

Happy coding

Introduction

Hey there, and welcome to Learn WordPress.

In this tutorial, you’re going to learn how to secure your WordPress plugins.

You will learn the benefits of ensuring your plugin code is secure, what steps to follow to secure your plugins, and where to find more information around plugin security.

What is plugin security?

Plugin security is the process of ensuring your plugin code not only works, but also does not introduce any security vulnerabilities. If your plugin has security vulnerabilities, it could make any WordPress site that has your plugin installed open to potential attacks, and can lead to that site being compromised.

When developing, it’s important to develop a security mindset, and to think about how your code could be used maliciously.

• Don’t trust any data, whether it’s user input, third party API data, or even data in your database. Always be checking to make sure it’s valid and safe to use.
• WordPress has a number of APIs that can help you with common tasks, such as sanitizing user input, validating data, and escaping output. Rely on using these APIs to help validate and sanitize your data instead of writing your own functions.
• Keep up to date with common vulnerabilities and keep your code up to date to prevent them.

Sanitizing inputs

One of the first steps to take when developing a plugin is to ensure that any user input is sanitized. This means that any data that is being passed to your plugin from the user, such as a form submission, or a URL parameter, is checked to make sure it’s safe to use.

In this example code, the name and email fields are being posted from a form that the plugin generates, and then saved in a custom database table called form_submissions;

$name = $_POST['name'];
$email = $_POST['email'];
global $wpdb;
$table_name = $wpdb->prefix . 'form_submissions';

$sql = "INSERT INTO $table_name (name, email) VALUES ('$name', '$email')";
$result = $wpdb->query($sql);

As you can see the data is being saved directly to the database, without any sanitization.

This means that if a user were to submit a name of John'; DROP TABLE form_submissions;-- the SQL INSERT query would be run, followed by the DROP query, and the table would be deleted from the database!

WordPress has a sanitization API that can be used to sanitize incoming data. You can use the sanitize_text_field and sanitize_email functions to sanitize the name and email fields before they’re used in the query.

$name = sanitize_text_field( $_POST['name'] );
$email = sanitize_email( $_POST['email'] );

global $wpdb;
$table_name = $wpdb->prefix . 'form_submissions';

$sql = "INSERT INTO $table_name (name, email) VALUES ('$name', '$email')";
$result = $wpdb->query($sql);

Notice that the code follows a key principle of sanitizing data, in that you do so as early as possible.

To read more about the available sanization functions available to WordPress developers, check out the Sanitizing page in the WordPress Developer Documentation.

Validating Data

Validating data is the process of testing it against a predefined pattern (or patterns) with a definitive result, either valid or invalid.

Untrusted data can come from many sources, users, third party API data, even your database data can be considered untrusted, especially if another plugin has modified it. Even site admins can make a mistake and enter incorrect or insecure data, so it’s important to always be checking your data.

In this example, a deletion function requires that a numeric ID is posted to an admin-ajax callback:

add_action( 'wp_ajax_delete_form_submission', 'wp_learn_delete_form_submission' );
function wp_learn_delete_form_submission() {
    if ( ! isset( $_POST['id'] ) ) {
        wp_send_json_error( 'Invalid ID' );
    }
    $id = $_POST['id'];

    global $wpdb;
    $table_name = $wpdb->prefix . 'form_submissions';

    $sql    = "DELETE FROM $table_name WHERE id = $id";
    $result = $wpdb->get_results( $sql );

    return wp_send_json( array( 'result' => $result ) );
}

Here the ID is being used directly in the SQL query, without any validation. Again, this means that if a user were to submit an ID of 1; DROP TABLE form_submissions; the same SQL DROP query would be run after the delete, and the table would be dropped.

To prevent this, you can use PHP’s type casting functionality, to ensure that the value of $id is always an integer value. This can be done by adding (int) before the variable name, like so:

$id = (int) $_POST['id'];

Note that this will only work if the first character in the string passed via the $_POST array can be cast as an integer, otherwise the value if $id will be 0. In that case it’s a good idea to update the code to handle this case.

Notice that code follows a key principle of validating data, in that you do so as early as possible.

if ($id === 0){
    // return early with an error
    return wp_send_json( array( 'result' => 'Invalid ID passed' ) );

}

To read more about the various ways to validate data in your WordPress plugin, check out the section on Validating Data in the WordPress Developer Documentation.

Escaping outputs

Another aspect of plugin security is to ensure that any information you output to the browser is safe, including any text, HTML or JavaScript code, or data from the database. Even if your plugin is not responsible for the source of the data being displayed, it is responsible for displaying it safely.

In this example, the code fetches the form submissions from the database, and then loops through the submissions and outputs the submission data in an admin screen in the WordPress dashboard:

    $submissions = wp_learn_get_form_submissions();
    ?>
    <div class="wrap" id="wp_learn_admin">
        <h1>Admin</h1>
        <table>
            <thead>
                <tr>
                    <th>Name</th>
                    <th>Email</th>
                </tr>
            </thead>
            <?php foreach ($submissions as $submission){ ?>
                <tr>
                    <td><?php echo $submission->name; ?></td>
                    <td><?php echo $submission->email; ?></td>
                    <td><a class="delete-submission" data-id="<?php echo $submission->id?>" style="cursor:pointer;">Delete</a></td>
                </tr>
            <?php } ?>
        </table>
    </div>
    <?php

Here we have three pieces of data that need to be escaped, the $submission->name and $submission->email fields, as well as the $submission->id.

<td><a class="delete-submission" data-id="<?php echo (int) $submission->id?>" style="cursor:pointer;">Delete</a></td>

For the name and email fields, we can use the built-in WordPress escaping function esc_html(). The ID can be escaped by casting it to an integer, as you might do for data validation.

  <td><?php echo esc_html( $submission->name ); ?></td>
  <td><?php echo esc_html( $submission->email ); ?></td>
  <td><a class="delete-submission" data-id="<?php echo (int) $submission->id?>" style="cursor:pointer;">Delete</a></td>

Notice that this code follows a key principle of escaping data, in that you escape the data as late as possible.

To read more about the various ways to escape outputs, check out the section on Escaping outputs in the WordPress Developer Documentation.

Preventing invalid requests

Whenever a request is made to functionality that your plugin provides, it’s important to check that the request is valid. This means checking that the request is coming from a trusted source.

For example, in your plugin code you might have a shortcode that renders a form, where users can submit their information. The function to render the form might look like this:

add_shortcode( 'wp_learn_form_shortcode', 'wp_learn_form_shortcode' );
function wp_learn_form_shortcode() {
    ob_start();
    ?>
    <form method="post">
        ?>
        <input type="hidden" name="wp_learn_form" value="submit">
        <div>
            <label for="email">Name</label>
            <input type="text" id="name" name="name" placeholder="Name">
        </div>
        <div>
            <label for="email">Email address</label>
            <input type="text" id="email" name="email" placeholder="Email address">
        </div>
        <div>
            <input type="submit" id="submit" name="submit" value="Submit">
        </div>
    </form>
    <?php
    $form = ob_get_clean();
    return $form;
}
?>

When the user submits the form, the data is sent to the server. The data is then processed by the plugin, sanitized, and stored in the database.

add_action( 'wp', 'wp_learn_maybe_process_form' );
function wp_learn_maybe_process_form() {
    if (!isset($_POST['wp_learn_form'])){
        return;
    }

    $name = sanitize_text_field( $_POST['name'] );
    $email = sanitize_email( $_POST['email'] );

    global $wpdb;
    $table_name = $wpdb->prefix . 'form_submissions';

    $sql = "INSERT INTO $table_name (name, email) VALUES ('$name', '$email')";
    $result = $wpdb->query($sql);
    if ( 0 < $result ) {
        wp_redirect( WPLEARN_SUCCESS_PAGE_SLUG );
        die();
    }

    wp_redirect( WPLEARN_ERROR_PAGE_SLUG );
    die();
}

Because the form might appear on any page where the shortcode is used, it’s possible that a malicious user could attempt to send a POST request to the form, either looking for a vulnerability in the plugin, or by sending multiple requests to the form.

To prevent this, you can check that the request is coming from a trusted source. To do this, you can implement something called a nonce, or a number used once.

First, in the form itself, you can add a nonce field by using the wp_nonce_field function, and passing a nonce action and nonce name to the function:

wp_nonce_field( 'wp_learn_form_nonce_action', 'wp_learn_form_nonce_field' );

When the form is rendered on the front end, a hidden field is added to the form, using the nonce name as the id and name attribute of the hidden field, and the generated nonce as the field value. This data is posted when the form is submitted.

Then, in the function that processes the form data, you can verify that the nonce is valid by using the wp_verify_nonce function, and by passing the POSTed nonce field and the nonce action to this function. If the result of this verification is false, you can exit early, preventing any further code execution.

    if ( wp_verify_nonce( $_POST['wp_learn_form_nonce_field'], 'wp_learn_form_nonce_action' ) {
        wp_redirect( WPLEARN_ERROR_PAGE_SLUG );
        die();
    }

Any time your plugin code makes a web request, be it via redirection to a new URL, POSTing data to a form, or making an AJAX request, you should check that the request is valid.

To read more about how to use nonces in your plugins, check out the section on Nonces in the WordPress Developer Documentation.

Preventing unauthenticated users

Depending on your plugin’s functionality, it’s a good idea to restrict certain features only to users with a specific permission level. For example, your plugin may have a function that deletes data from the database.

function wp_learn_delete_form_submission() {
    if ( ! isset( $_POST['id'] ) ) {
        wp_send_json_error( 'Invalid ID' );
    }
    $id = (int) $_POST['id'];

    global $wpdb;
    $table_name = $wpdb->prefix . 'form_submissions';

    $sql    = "DELETE FROM $table_name WHERE id = $id";
    $result = $wpdb->get_results( $sql );

    return wp_send_json( array( 'result' => $result ) );
}

While you might do your best to prevent anyone who doesn’t have permission to run this function, it’s still a good idea to include checks against this situation.

WordPress includes a robust User Roles and Capabilities system, which allows you to either use the default user roles and capabilities, or create custom ones.

In this case, it could be as easy as only allowing users who have the capability to manage site options, which is a standard capability that is included with the administrator role.

function wp_learn_delete_form_submission() {
    if ( ! current_user_can( 'manage_options' ) ) {
        return wp_send_json( array( 'result' => 'Authentication error' ) );
    }
    // rest of function code
}

The WordPress Developer Documentation has a detailed section on User Roles and Capabilities, which includes a list of the default capabilities, and how to create custom ones.

Finally, make sure to read the entry in the WordPress Developer Documentation on Security, as it includes all the examples in this video, as well as additional information on security best practices, common vulnerabilities, and more example code.

Happy coding!