agentleFS
wpultimatesecuritywpultimatesecurity/WordPress-Security-Skills/skills/object-injection-deserialization/SKILL.md

object-injection-deserialization

Skill45 starsChanged 54 days agoMITView on GitHub

Use when code calls unserialize, maybe_unserialize, or stores serialized PHP in options, meta, or transients from untrusted input. Avoids unserialize on attacker-controlled data, prefers json_encode / json_decode, and when unserialize is unavoidable passes ['allowed_classes' => false]. Prevents PHP object injection and POP-chain remote code execution.

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill whenever code touches PHP serialization:

    line 19

≈ “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

Object injection & deserialization security

When to use this skill

Use this skill whenever code touches PHP serialization:

  • Calling unserialize(), maybe_unserialize(), or is_serialized() on user input.
  • Storing serialized arrays/objects in options, post meta, user meta, or transients.
  • Importing data from files, feeds, or uploads that may contain serialized PHP.
  • Migrating or caching objects by serializing them.

PHP object injection lets an attacker instantiate arbitrary objects and trigger magic methods (__wakeup(), __destruct(), __toString()) that form POP (property-oriented programming) chains, often leading to remote code execution.

Core principles (and why they matter)

  1. Never unserialize() untrusted input. Anything from $_POST, $_GET, files, remote APIs, or cookies is untrusted.
  2. Prefer JSON for interchange. wp_json_encode() / json_decode() only produces arrays/stdClass and cannot instantiate PHP classes, breaking object-injection chains.
  3. maybe_unserialize() is still unserialize(). It just checks is_serialized() first. Do not use it on attacker-controlled data.
  4. If forced to unserialize, disable classes. Pass array( 'allowed_classes' => false ) (PHP 7.0+) to unserialize() so no objects are instantiated.
  5. Do not trust is_serialized() as a safety check. It only tells you the format; it does not make the payload safe.
  6. Keep serialized objects out of user-facing storage. Options and meta that users can influence should store JSON or scalar values.

Step-by-step implementation

  1. Identify every unserialize(), maybe_unserialize(), and is_serialized() call.
  2. Determine whether the data is attacker-controllable. If yes, replace with JSON.
  3. For trusted internal serialization, use allowed_classes => false or a narrow allowlist of classes.
  4. Validate the decoded shape (expected keys/types) before use.
  5. Escape any values from the decoded data before output.

Supporting references

ReferenceLoad when
Object injection & deserialization checklistBefore final verification of the object injection & deserialization controls.
Secure deserialization and importImplementing validated JSON imports and disabling class instantiation where trusted deserialization is unavoidable.

Common AI mistakes / anti-patterns

Mistake 1 — unserialize() on $_POST

// ❌ Insecure: arbitrary object instantiation.
$data = unserialize( $_POST['config'] );
// ✅ Secure: use JSON for user-supplied structured data.
$raw  = isset( $_POST['config'] ) ? wp_unslash( $_POST['config'] ) : '';
$data = json_decode( $raw, true );
if ( ! is_array( $data ) ) {
    wp_die( esc_html__( 'Invalid configuration.', 'my-plugin' ), 400 );
}
$api_key = isset( $data['api_key'] ) ? sanitize_text_field( $data['api_key'] ) : '';

Mistake 2 — Treating maybe_unserialize() as safe

// ❌ Insecure: maybe_unserialize is unserialize with a format check.
$config = maybe_unserialize( file_get_contents( $uploaded_file ) );
// ✅ Secure: read and validate JSON instead.
$raw      = file_get_contents( $uploaded_file ); // phpcs:ignore WordPress.WP.AlternativeFunctions.file_get_contents_file_get_contents
$decoding = json_decode( $raw, true );
if ( null === $decoding && json_last_error() !== JSON_ERROR_NONE ) {
    wp_die( esc_html__( 'Invalid import file.', 'my-plugin' ), 400 );
}

Mistake 3 — Trusting is_serialized()

// ❌ Insecure: is_serialized only describes the format.
if ( is_serialized( $value ) ) {
    $value = unserialize( $value ); // still dangerous
}
// ✅ Secure: if you must accept serialized data, disable class instantiation.
if ( is_serialized( $value ) ) {
    $value = @unserialize( $value, array( 'allowed_classes' => false ) );
}

Mistake 4 — Storing user-supplied serialized blobs

// ❌ Insecure: stores attacker-controlled serialized payload in meta.
update_post_meta( $post_id, '_my_plugin_data', $_POST['data'] );
// ✅ Secure: store JSON; validate and sanitize on read.
$raw  = isset( $_POST['data'] ) ? wp_unslash( $_POST['data'] ) : '';
$data = json_decode( $raw, true );
if ( is_array( $data ) ) {
    update_post_meta( $post_id, '_my_plugin_data', wp_json_encode( $data ) );
}

Mistake 5 — Serializing objects for cache/transients without class restrictions

// ❌ Risky: transient value can instantiate arbitrary objects on read.
set_transient( 'my_plugin_state', serialize( $object ) );
// ✅ Safer: store JSON or, if objects are required, a known shape with allowed_classes false.
set_transient( 'my_plugin_state', wp_json_encode( $object ) );

Correct code examples

A safe importer that rejects serialized PHP and parses JSON is in references/secure-deserialization.php.

Checklist

  • No unserialize() / maybe_unserialize() runs on user input, uploads, or remote data.
  • JSON is used for structured data interchange.
  • Where unserialize() is unavoidable, allowed_classes => false is set.
  • is_serialized() is not treated as a security check.
  • Serialized blobs are not stored from user-controlled sources.
  • Decoded JSON is validated for expected shape before use.
  • Values from decoded data are sanitized/escaped before output.

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.