WP-CLI security
SkillSecurityUse when registering a WP-CLI command with WP_CLI::add_command or writing command logic. Validates and sanitizes positional and associative arguments, does not assume a logged-in user or capability context, avoids printing secrets, and confirms destructive operations. Prevents injection and unsafe automation through the CLI surface.
Available today. Use it from your connected AI after setup.
No other account needed.
Connect ahel once, and every AI you use reads what you have installed.
Then ask your AI: use the WP-CLI security skill
What this skill tells your AI
The instructions your AI receives, as published by wpultimatesecurity/wordpress-security-skills in skills/wp-cli-security/SKILL.md and read by ahel’s review.
When to use this skill
Use this skill whenever code registers or implements a WP-CLI command:
WP_CLI::add_command( 'my-plugin thing', ... ).- Writing a command class method that reads
$args/$assoc_args. - Running destructive operations (delete, reset, import, purge) from the command line.
- Returning data or logging progress from a CLI routine.
WP-CLI runs as the system user, often as root on containers or with broad file-system
permissions. It has no WordPress current user by default, so current_user_can()
does not authorize CLI actions. Arguments are still untrusted input.
Related: see the capability-permission-checks skill for normal web-request
authorization and the cron-background-job-security skill for scheduled jobs.
Core principles (and why they matter)
- No
current_user_can()in CLI by default. There is no logged-in WordPress user. Either useWP_CLI::launch_self()with a--userflag, explicitly set a user with--user, or perform capability-independent maintenance tasks only. - Sanitize every argument.
$argsand$assoc_argscan come from scripts, CI, or shell history; treat them as untrusted input. - Use prepared statements for database access. Do not interpolate
$argsinto$wpdbqueries. - Confirm destructive actions.
WP_CLI::confirm()stops accidental data loss in automated runs. - Never print secrets. API keys, tokens, and passwords must not appear in CLI output, logs, or shell history.
- Be explicit about
--allow-root. Running as root is common in containers but dangerous; document it and avoid file ownership surprises.
Step-by-step implementation
- Register the command with
WP_CLI::add_command()and a clear docblock. - In the command method, validate argument counts and sanitize each value to type.
- For destructive commands, call
WP_CLI::confirm()unless--yesis passed. - Use
$wpdb->prepare()for any dynamic query. - Avoid echo; use
WP_CLI::log(),WP_CLI::success(), orWP_CLI::error(). - If the command needs a user context, require
--user=<id|login|email>and load it withWP_User::get_data_by()/wp_set_current_user().
Supporting references
| Reference | Load when |
|---|---|
| WP-CLI security checklist | Before final verification of the wp-cli security controls. |
| Secure WP-CLI command | Implementing a WP-CLI command with sanitized arguments, prepared queries, destructive-action confirmation, and explicit user context. |
Common AI mistakes / anti-patterns
Mistake 1 — Interpolating $args into a query
// ❌ Insecure: SQL injection via CLI argument.
$wpdb->query( "DELETE FROM {$wpdb->prefix}my_table WHERE id = {$args[0]}" );
// ✅ Secure: use $wpdb->prepare().
$id = absint( $args[0] );
$wpdb->query( $wpdb->prepare( "DELETE FROM {$wpdb->prefix}my_table WHERE id = %d", $id ) );
Mistake 2 — Assuming an admin user context
// ❌ Insecure: current_user_can is meaningless without --user.
if ( current_user_can( 'manage_options' ) ) {
update_option( 'my_plugin_key', $value );
}
// ✅ Secure: require --user or gate by filesystem/CLI policy, not web capability.
class My_Plugin_CLI_Command {
/**
* Update a setting.
*
* ## OPTIONS
* [--user=<user>]
* : User to run as.
*/
public function update_setting( $args, $assoc_args ) {
if ( isset( $assoc_args['user'] ) ) {
$user = get_user_by( 'login', $assoc_args['user'] );
if ( $user ) {
wp_set_current_user( $user->ID );
}
}
if ( ! current_user_can( 'manage_options' ) ) {
WP_CLI::error( 'This command requires --user with manage_options capability.' );
}
update_option( 'my_plugin_key', sanitize_text_field( $args[0] ) );
WP_CLI::success( 'Setting updated.' );
}
}
Mistake 3 — Printing secrets in output
// ❌ Insecure: secret is now in shell history and terminal scrollback.
WP_CLI::log( 'Using API key: ' . $api_key );
// ✅ Secure: log an identifier or redacted value.
WP_CLI::log( 'Using API key ending in: ' . substr( $api_key, -4 ) );
Mistake 4 — Destructive command with no confirmation
// ❌ Risky: a typo in CI wipes data.
public function purge( $args ) {
$wpdb->query( "TRUNCATE {$wpdb->prefix}my_plugin_logs" );
}
// ✅ Secure: require confirmation unless --yes is passed.
public function purge( $args, $assoc_args ) {
if ( ! isset( $assoc_args['yes'] ) ) {
WP_CLI::confirm( 'This will delete all logs. Are you sure?' );
}
global $wpdb;
$wpdb->query( "TRUNCATE {$wpdb->prefix}my_plugin_logs" );
WP_CLI::success( 'Logs purged.' );
}
Mistake 5 — Trusting associative args without defaults
// ❌ Fragile: missing key causes undefined array warning.
$days = $assoc_args['days'];
// ✅ Secure: provide defaults and sanitize.
$days = isset( $assoc_args['days'] ) ? absint( $assoc_args['days'] ) : 30;
Correct code examples
A complete secure WP-CLI command class is in
references/secure-wp-cli-command.php.
Checklist
- Every CLI argument is sanitized to its expected type.
- Database queries use
$wpdb->prepare(); no argument interpolation. - Destructive commands call
WP_CLI::confirm()unless--yesis provided. - The command does not rely on
current_user_can()without an explicit--user. - Secrets, tokens, and API keys are never printed or logged in plaintext.
-
--allow-rootusage is documented and justified. - Output uses
WP_CLI::log()/success()/error()instead ofecho. - Command docblocks define options and examples.
Official references
Signals
- GitHub stars
- 31
- Forks
- 2
- Last commit
- Sep 2026
Advanced
- Catalog kind
- skill
- Gateway key
wp-cli-security- Source
- github.com/wpultimatesecurity/wordpress-security-skills