agentleFS
wpultimatesecuritywpultimatesecurity/WordPress-Security-Skills/skills/wp-hardening-best-practices/SKILL.md

wp-hardening-best-practices

Skill45 starsChanged 54 days agoMITView on GitHub

Use when configuring or hardening a WordPress site, editing wp-config.php, writing .htaccess or nginx rules, setting file permissions, or advising on deployment security. Covers security keys, DISALLOW_FILE_EDIT, FORCE_SSL_ADMIN, disabling debug output, blocking PHP execution in uploads, protecting sensitive files, and least-privilege file permissions. Apply proactively when setting up or reviewing a site's configuration.

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 3 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill for site/environment configuration, as opposed to plugin code:

    line 19
  2. Very specific:

    Place wp-config.php permissions at 640/600.

    line 54
  3. Very specific:

    Add server rules to deny PHP in uploads, protect sensitive files, and (optionally) restrict wp-admin/xmlrpc.php.

    line 56
  4. Distinctive:

    Keep everything updated; remove what you don't use.

    line 59
  5. Distinctive:

    Add abuse controls for public surfaces: comment moderation on, pingbacks off if unused, and — for bot-heavy sites — CAPTCHA/Turnstile on …

    line 60

≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

WordPress hardening best practices

When to use this skill

Use this skill for site/environment configuration, as opposed to plugin code:

  • Editing wp-config.php (keys, constants, debug).
  • Writing/reviewing .htaccess (Apache) or server-block (nginx) rules.
  • Setting filesystem permissions and ownership.
  • Locking down wp-admin, XML-RPC, REST user enumeration, and the uploads dir.
  • Pre-launch hardening checklists and deployment review.

This complements the secure-coding skills: even perfect code runs on a host that must be configured to fail safely.

Core principles (and why they matter)

  1. Disable in-dashboard file editing in production. DISALLOW_FILE_EDIT removes the plugin/theme editor — a single compromised admin session otherwise becomes code execution.
  2. Force TLS for admin and logins. FORCE_SSL_ADMIN stops credentials and cookies traveling in cleartext.
  3. Never display errors in production. Stack traces leak paths, queries, and secrets. WP_DEBUG off, display_errors off; log to a non-web-readable file if needed.
  4. Block PHP execution where only data should live. An attacker who slips a PHP file into wp-content/uploads gets RCE unless the server refuses to execute it there.
  5. Protect sensitive files. wp-config.php, .htaccess, readme.html, debug.log, and dotfiles should not be web-readable.
  6. Least privilege on the filesystem. Files 644, directories 755, wp-config.php 640/600; web server should not own files it doesn't need to write.
  7. Unique, strong security keys/salts. They invalidate stolen cookies; regenerate if leaked.
  8. Reduce attack surface. Disable XML-RPC if unused, block user enumeration, keep core/ plugins/themes updated, remove unused plugins.

Step-by-step implementation

  1. In wp-config.php: set unique salts, DISALLOW_FILE_EDIT, FORCE_SSL_ADMIN, disable debug display; optionally WP_AUTO_UPDATE_CORE, DISALLOW_FILE_MODS for locked builds.
  2. Place wp-config.php permissions at 640/600; ensure it is above or protected from the web root.
  3. Add server rules to deny PHP in uploads, protect sensitive files, and (optionally) restrict wp-admin/xmlrpc.php.
  4. Set file/dir permissions to least privilege.
  5. Keep everything updated; remove what you don't use.
  6. Add abuse controls for public surfaces: comment moderation on, pingbacks off if unused, and — for bot-heavy sites — CAPTCHA/Turnstile on login and registration via an established plugin (hand-rolled CAPTCHAs fail in both directions). Pair with the login throttle from authentication-session-security; XML-RPC stays off unless needed.

Supporting references

ReferenceLoad when
WordPress hardening checklistBefore final verification of the wordpress hardening controls.
Pre-launch (go-live) checklistReviewing a production launch across configuration, code, data, and operations.
Apache and nginx hardening rulesImplementing server hardening for sensitive files, uploads, and restricted endpoints.
Hardened wp-config.php constantsImplementing the wp-config.php hardening flow for salts, HTTPS, debugging, editing, and updates.

Common AI mistakes / anti-patterns

Mistake 1 — Leaving debug output on in production

// ❌ Insecure: errors rendered to visitors leak paths, SQL, secrets.
define( 'WP_DEBUG', true );
define( 'WP_DEBUG_DISPLAY', true );
// ✅ Secure: log privately, never display, in production.
define( 'WP_DEBUG', false );
// If you must debug on a live box, at least never display:
@ini_set( 'display_errors', '0' );
define( 'WP_DEBUG_DISPLAY', false );
define( 'WP_DEBUG_LOG', true ); // writes to wp-content/debug.log — block it from the web

Mistake 2 — Allowing the dashboard file editor

// ❌ Risky: compromised admin = arbitrary PHP via Appearance/Plugins editor.
// (default: editor enabled)
// ✅ Secure: disable file editing (and optionally all file mods).
define( 'DISALLOW_FILE_EDIT', true );
// Fully locked deploy (no plugin/theme install/update via UI):
define( 'DISALLOW_FILE_MODS', true );

Mistake 3 — 777 permissions "to make it work"

# ❌ Insecure: world-writable lets any local process modify your site.
chmod -R 777 wp-content
# ✅ Secure: least privilege.
find . -type d -exec chmod 755 {} \;
find . -type f -exec chmod 644 {} \;
chmod 640 wp-config.php

Mistake 4 — Uploads directory that executes PHP

# ❌ Insecure: a smuggled shell.php in uploads runs.
# (no restriction)
# ✅ Secure: deny PHP execution in uploads (Apache). Place in wp-content/uploads/.htaccess
<FilesMatch "\.(?i:php|php\d|phtml|phar)$">
    Require all denied
</FilesMatch>

Mistake 5 — Default/duplicated security keys

// ❌ Insecure: placeholder salts (or copied between sites).
define( 'AUTH_KEY', 'put your unique phrase here' );
// ✅ Secure: generate unique values from the official salt API and rotate if leaked.
// https://api.wordpress.org/secret-key/1.1/salt/
define( 'AUTH_KEY', '...64 random chars...' );
// (all eight: AUTH_KEY/SALT, SECURE_AUTH_KEY/SALT, LOGGED_IN_KEY/SALT, NONCE_KEY/SALT)

Mistake 6 — Leaving XML-RPC enabled when unused

// ❌ Risky: XML-RPC is a brute-force and pingback amplification vector if not needed.
// (default: enabled)
// ✅ Secure: disable XML-RPC entirely if the site does not need it.
add_filter( 'xmlrpc_enabled', '__return_false' );
// Or block at the server level (see htaccess-hardening.conf).

Mistake 7 — Allowing REST user enumeration

# ❌ Risky: /wp-json/wp/v2/users/ reveals usernames to unauthenticated visitors.
curl https://example.com/wp-json/wp/v2/users/
// ✅ Safer: require authentication for the users endpoint.
add_filter( 'rest_endpoints', function ( $endpoints ) {
    if ( isset( $endpoints['/wp/v2/users'] ) ) {
        $endpoints['/wp/v2/users'][0]['permission_callback'] = static function () {
            return is_user_logged_in();
        };
    }
    return $endpoints;
} );

Mistake 8 — Application Passwords enabled without review

Application Passwords (WordPress 5.6+) are powerful for integrations but create a long-lived credential surface. Disable them if unused, or restrict them to users who actually need machine access.

// ✅ Secure: disable Application Passwords if the site does not use them.
add_filter( 'wp_is_application_passwords_available', '__return_false' );

Related: see the secrets-credentials-management skill for storing and handling API keys and tokens safely. See the security-headers-csp skill for application-level security headers (HSTS belongs in this skill's server config; the rest is code-level), and the dependency-supply-chain-security skill for keeping bundled libraries and CDN assets from rotting.

Correct code examples

Hardened wp-config.php constants are in references/wp-config-hardening.php; Apache and nginx rules (deny PHP in uploads, protect wp-config.php/.htaccess/debug.log, limit xmlrpc.php) are in references/htaccess-hardening.conf; the consolidated pre-launch sweep is references/go-live-checklist.md.

Checklist

  • Unique, strong security keys/salts set (all eight).
  • DISALLOW_FILE_EDIT enabled in production.
  • FORCE_SSL_ADMIN enabled; site served over HTTPS.
  • WP_DEBUG/WP_DEBUG_DISPLAY off in production; logs not web-readable.
  • PHP execution denied in wp-content/uploads.
  • wp-config.php, .htaccess, debug.log, dotfiles not web-accessible.
  • Permissions least-privilege: dirs 755, files 644, wp-config.php 640/600.
  • XML-RPC disabled/limited if unused; user enumeration limited.
  • Core, plugins, themes kept updated; unused extensions removed.
  • Admin accounts use strong passwords + 2FA where possible.

Official references

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.