README

Laravel Scrubber

A Laravel package to scrub sensitive information that breaks operational security policies from being leaked on accident ~~or not~~ by developers.

Requirements

PHP 8.2, 8.3, 8.4, or 8.5
Laravel 10.x, 11.x, or 12.x

Installation

install the package via composer:

composer require yorcreative/laravel-scrubber

Publish the packages assets.

php artisan vendor:publish --provider="YorCreative\Scrubber\ScrubberServiceProvider"

Configuration

Adjust the configuration file to suite your application, located in /config/scrubber.php.

return [
    /**
     * Specify the string to use to redact the data
     */
    'redaction' => '**redacted**',

    'secret_manager' => [
        'key' => env('APP_KEY'),
        'cipher' => 'AES-256-CBC',
        'enabled' => false,
        'providers' => [
            // See "Secret Manager Providers" section for full configuration options
            'gitlab' => ['enabled' => false, /* ... */],
            'aws' => ['enabled' => false, /* ... */],
            'vault' => ['enabled' => false, /* ... */],
            'azure' => ['enabled' => false, /* ... */],
            'google' => ['enabled' => false, /* ... */],
        ],
    ],

    /**
     * Specify the regexes to load
     * You can use a wildcard (*) to load all regexes in all `custom_regex_namespaces` and the default core regexes.
     * Otherwise, specify the regexes you want to load either by qualified class name or by unqualified (base) class name,
     * which will then search the `custom_regex_namespaces` and the default core regexes for a match.
     */
    'regex_loader' => ['*'],
    
    /**
     * Specify regex patterns to exclude from loading when using the regex loader
     * This allows fine-grained control over which regex patterns are loaded, especially useful when using wildcard (*) in regex_loader
     * 
     * You can exclude patterns using any of these formats:
     * - Fully qualified class name (e.g., 'YorCreative\Scrubber\RegexCollection\GoogleApi')
     * - Base class name (e.g., 'GoogleApi', 'EmailAddress')
     * - Pattern constant from RegexCollection (e.g., RegexCollection::$GOOGLE_API)
     * - Custom namespace class (e.g., 'App\Scrubber\RegexCollection\HerokuApiKey')
     * 
     * Example:
     * [
     *     'GoogleApi',
     *     'YorCreative\Scrubber\RegexCollection\EmailAddress',
     *     RegexCollection::$HEROKU_API_KEY,
     *     'App\Scrubber\RegexCollection\HerokuApiKey'
     * ]
     */
    'exclude_regex' => [],


    /**
     * Specify namespaces from which regexes will be loaded when using the wildcard (*)
     * for the regex_loader or where you use unqualified class names.
     */
    'custom_regex_namespaces' => [
       'App\\Scrubber\\RegexCollection',
    ],

    /**
     * Specify config keys for which the values will be scrubbed
     * You should use the dot notation to specify the keys
     * You can use wildcards (*) to match multiple keys
     *
     *  - 'database.connections.*.password'
     *  - 'app.secrets.*'
     *  - 'app.some.nested.key'
     */
    'config_loader' => [
        '*token',
        '*key',
        '*secret',
        '*password',
    ],

    /**
     * Minimum character length for config values to be treated as scrubbable.
     * Values shorter than this will be ignored to prevent overly aggressive
     * scrubbing (e.g., Livewire's release_token defaults to 'a').
     */
    'config_loader_min_length' => 4,

    /**
     * Config key patterns to exclude from scrubbing.
     * Supports wildcards (*) via Str::is().
     */
    'config_loader_exclusions' => [],

    /**
     * Specify the channels to tap into
     * You can use wildcards (*) to match multiple channels
     */ 
    'tap_channels' => false,
];

Usage

The scrubber can be utilized in two ways, the first one being a Log scrubber. A tap is added to detect and sanitize any sensitive information from hitting a log file. The second way is to integrate into your application and utilize the Scrubber directly. This way is particular useful if you, for example, would like to detect and sanitize any messages on a messaging platform.

Logging Detection & Sanitization

Log::info('some message', [
    'context' => 'accidental',
    'leak_of' => [
        'jwt' => '<insert jwt token here>'
    ]
])

// testing.INFO: some message {"context":"accidental","leak_of":{"jwt": '**redacted**'}} 

Log::info('<insert jwt token here>')

// testing.INFO: **redacted**

Direct Usage for Detection & Sanitization

Scrubber::processMessage([
    'context' => 'accidental',
    'leak_of' => [
        'jwt' => '<insert jwt token here>'
    ]
]);
// [
//     "context" => "accidental"
//     "leak_of" => [
//         "jwt" => "**redacted**"
//     ]
// ];

Scrubber::processMessage('<insert jwt token here>');
// **redacted**

Detection Statistics API

Track what patterns are matching and how often:

// Get scrubbing statistics for the current request
$stats = Scrubber::getStats();
// ['total_scrubs' => 5, 'patterns_matched' => ['JsonWebToken' => 2, 'EmailAddress' => 3]]

// Test a string without modifying stats - useful for debugging
$result = Scrubber::test('Contact: john@example.com, SSN: 123-45-6789');
// [
//     'matched' => true,
//     'patterns' => ['EmailAddress' => 1, 'SocialSecurityNumber' => 1],
//     'scrubbed' => 'Contact: **redacted**, SSN: ***-**-****'
// ]

// Reset statistics between requests
Scrubber::resetStats();

Events

The scrubber can dispatch a SensitiveDataDetected event each time a pattern matches during scrubbing. This is useful for alerting, metrics, or audit logging.

Enable events in your config:

'events' => [
    'enabled' => true,
],

The event carries three public properties:

Property	Type	Description
`patternName`	`string`	The class basename of the matched pattern (e.g. `JsonWebToken`)
`hitCount`	`int`	Number of matches found for that pattern in the content
`context`	`string`	Either `log` (triggered via log tap) or `manual` (triggered via `Scrubber::processMessage()`)

use YorCreative\Scrubber\Events\SensitiveDataDetected;

Event::listen(SensitiveDataDetected::class, function (SensitiveDataDetected $event) {
    // $event->patternName  — e.g. 'JsonWebToken'
    // $event->hitCount     — e.g. 2
    // $event->context      — 'log' or 'manual'
});

Log Channel Opt-in

This package provides you the ability to define through the configuration file what channels you want to scrub specifically. By default, this package ships with a wildcard value and opts in to scrub all the log channels in your application.

Defining Log Channel Opt-in

To opt in to one or more channels, list the channel(s) name into the tap_channels array in the config.

'tap_channels' => [
    'single',
    'papertrail'
]

To disable tap logging functionality and use the package independently and not tap your Laravel application logging, modify the config file by setting the tap_channels field as follows:

'tap_channels' => false

Regex Class Opt-in

You have the ability through the configuration file to define what regex classes you want loaded into the application when it is bootstrapped. By default, this package ships with a wildcard value.

Regex Collection & Defining Opt-in

To opt in, utilize the static properties on the RegexCollection class.

 'regex_loader' => [
        RegexCollection::$GOOGLE_API,
        RegexCollection::$AUTHORIZATION_BEARER,
        RegexCollection::$CREDIT_CARD_AMERICAN_EXPRESS,
        RegexCollection::$CREDIT_CARD_DISCOVER,
        RegexCollection::$CREDIT_CARD_VISA,
        RegexCollection::$JSON_WEB_TOKEN
    ],

Note: The package includes 31 built-in patterns. See all available patterns in RegexCollection.php.

PII Detection with Partial Masking

The following patterns use contextual replacement values for improved readability instead of the generic **redacted**:

Pattern	Detects	Masked Output
`RegexCollection::$SOCIAL_SECURITY_NUMBER`	US Social Security Numbers	`*--****`
`RegexCollection::$PHONE_NUMBER`	Phone numbers (US/International)	`(*) -***`
`RegexCollection::$IP_ADDRESS_V4`	IPv4 addresses	`*...*`
`RegexCollection::$IP_ADDRESS_V6`	IPv6 addresses	`**::**:...`
`RegexCollection::$IBAN`	International Bank Account Numbers	`********************`

Scrubber::processMessage('SSN: 123-45-6789, Phone: (555) 123-4567');
// "SSN: ***-**-****, Phone: (***) ***-****"

Scrubber::processMessage('Server IP: 192.168.1.1');
// "Server IP: ***.***.***.***"

Opting Into Custom Extended Classes

To create custom scrubbers, see the Extending the Scrubber section.

The regex_loader array takes strings, not objects. To opt in to specific custom extended regex classes, define the class name as a string.

For example if I have a custom extended class as such:

<?php

namespace App\Scrubber\RegexCollection;

use YorCreative\Scrubber\Interfaces\RegexCollectionInterface;

class TestRegex implements RegexCollectionInterface
{
    public function getPattern(): string
    {
        /**
         * @note return a regex pattern to detect a specific piece of sensitive data.
         */
        return '(?<=basic) [a-zA-Z0-9=:\\+\/-]{5,100}';
    }

    public function getTestableString(): string
    {
        /**
         * @note return a string that can be used to verify the regex pattern provided.
         */
        return 'basic f9Iu+YwMiJEsQu/vBHlbUNZRkN/ihdB1sNTU';
    }
    
    public function getReplacementValue(): string
    {
        
        /**
         * @note return a string that replaces the regex pattern provided.
         */
        return config('scrubber.redaction');
    }

    public function isSecret(): bool
    {
        return false;
    }
}

The regex_loader array should be defined as such:

 'regex_loader' => [
        RegexCollection::$GOOGLE_API,
        RegexCollection::$AUTHORIZATION_BEARER,
        RegexCollection::$CREDIT_CARD_AMERICAN_EXPRESS,
        RegexCollection::$CREDIT_CARD_DISCOVER,
        RegexCollection::$CREDIT_CARD_VISA,
        RegexCollection::$JSON_WEB_TOKEN,
        'TestRegex'
    ],

RegexCollection & Defining Opt-out

When using wildcard loading ('regex_loader' => ['*']), you can exclude specific regex patterns using the exclude_regex configuration. This allows you to load all patterns except those explicitly excluded.

'exclude_regex' => [
    // Exclude by base class name
    'GoogleApi',
    
    // Exclude by fully qualified class name
    'YorCreative\Scrubber\RegexCollection\EmailAddress',
    
    // Exclude using RegexCollection constant
    RegexCollection::$HEROKU_API_KEY,
    
    // Exclude from custom namespace
    'App\Scrubber\RegexCollection\HerokuApiKey'
],

The exclude_regex configuration supports multiple formats for excluding patterns:

Base class names (e.g., 'GoogleApi')
Fully qualified class names
RegexCollection constants
Custom namespace classes

This is particularly useful when you want to use most patterns but need to exclude a few specific ones from your scrubbing process.

Config Loader Filtering

The config_loader scrubs config values matching key patterns (e.g., *password, *token). Two additional options give you control over which values get scrubbed.

Minimum Length

By default, config values shorter than 4 characters are ignored. This prevents overly aggressive scrubbing when packages use short default values (e.g., Livewire's release_token defaults to 'a', which would cause every letter "a" in your logs to be redacted).

// Default: ignore values shorter than 4 characters
'config_loader_min_length' => 4,

// Set to 0 to disable the minimum length filter
'config_loader_min_length' => 0,

Non-string config values (booleans, integers, etc.) are always excluded regardless of this setting.

Key Exclusions

Certain config keys may match broad loader patterns but should not be treated as sensitive values. The config_loader_exclusions option allows you to bypass scrubbing for specific keys or key prefixes using wildcard patterns:

'config_loader_exclusions' => [
    // Exclude a specific key
    'livewire.release_token',

    // Exclude all keys under a namespace
    'livewire.*',
],

About the Scrubber

This package provides the ability to pull in secrets from external sources. Providing the ability to detect information leakage, and sanitize secrets without needing an exact regex pattern to detect it.

Encryption

For enhanced application security, all secrets pulled, from any provider, are encrypted and only decrypted to run the detection. You can see this in action here.

Secret Manager Providers

Laravel Scrubber supports pulling secrets from multiple external secret management services. This allows you to detect and sanitize secrets without needing exact regex patterns - if a value matches a secret from your vault, it gets scrubbed.

To enable secret managers, set secret_manager.enabled to true in your config and enable one or more providers.

GitLab CI/CD Variables

Pull secrets from GitLab project variables.

'gitlab' => [
    'enabled' => true,
    'project_id' => env('GITLAB_PROJECT_ID'),
    'token' => env('GITLAB_TOKEN'),
    'host' => 'https://gitlab.com', // Or your self-hosted GitLab URL
    'keys' => ['*'], // Or specific variable names: ['DB_PASSWORD', 'API_KEY']
],

See GitLab's documentation on adding project variables.

AWS Secrets Manager

Pull secrets from AWS Secrets Manager. Requires the AWS SDK.

composer require aws/aws-sdk-php

'aws' => [
    'enabled' => true,
    'region' => env('AWS_DEFAULT_REGION', 'us-east-1'),
    'version' => 'latest',
    'credentials' => [
        'key' => env('AWS_ACCESS_KEY_ID'),
        'secret' => env('AWS_SECRET_ACCESS_KEY'),
    ],
    'keys' => ['*'], // Or specific secret names/ARNs
],

The AWS SDK automatically uses the default credential chain (environment variables, IAM roles, ECS task roles, etc.) when credentials are not explicitly provided.

HashiCorp Vault

Pull secrets from HashiCorp Vault using REST API. No additional dependencies required.

'vault' => [
    'enabled' => true,
    'host' => env('VAULT_ADDR', 'http://127.0.0.1:8200'),
    'token' => env('VAULT_TOKEN'),
    'namespace' => env('VAULT_NAMESPACE'), // Enterprise feature (optional)
    'engine' => env('VAULT_ENGINE', 'secret'), // KV engine mount path
    'path' => env('VAULT_PATH', ''), // Base path within the engine
    'version' => env('VAULT_KV_VERSION', 2), // KV engine version (1 or 2)
    'keys' => ['*'], // Or specific secret paths
],

Azure Key Vault

Pull secrets from Azure Key Vault using REST API. No additional dependencies required.

'azure' => [
    'enabled' => true,
    'vault_url' => env('AZURE_VAULT_URL'), // https://my-vault.vault.azure.net
    // Authentication options (in order of precedence):
    // Option 1: Direct access token
    'access_token' => env('AZURE_VAULT_ACCESS_TOKEN'),
    // Option 2: Client credentials (service principal)
    'tenant_id' => env('AZURE_TENANT_ID'),
    'client_id' => env('AZURE_CLIENT_ID'),
    'client_secret' => env('AZURE_CLIENT_SECRET'),
    'keys' => ['*'], // Or specific secret names
],

Authentication methods (in order of precedence):

Direct access token - For testing or short-lived tokens
Managed Identity - Auto-detected when running in Azure (App Service, Functions, VMs)
Client credentials - Service principal with tenant_id, client_id, client_secret

Google Cloud Secret Manager

Pull secrets from Google Cloud Secret Manager using REST API. No additional dependencies required.

'google' => [
    'enabled' => true,
    'project_id' => env('GOOGLE_CLOUD_PROJECT'),
    'access_token' => env('GOOGLE_SECRET_MANAGER_TOKEN'), // Optional
    'keys' => ['*'], // Or specific secret names
],

Authentication methods (in order of precedence):

Direct access token - For testing or when running outside GCP
Application Default Credentials - Auto-detected when running in GCP (Compute Engine, Cloud Run, GKE, Cloud Functions)

When running on GCP, leave access_token empty to use ADC automatically.

JSON Secret Values

All providers support JSON-formatted secret values. Nested values are automatically flattened:

// Secret named "database-config" with value:
{
  "host": "db.example.com",
  "credentials": {
    "username": "admin",
    "password": "secret123"
  }
}

This creates three scrubber patterns:

database-config.host → db.example.com
database-config.credentials.username → admin
database-config.credentials.password → secret123

Extending the Scrubber

Creating new Scrubber Detection Classes

php artisan make:regex-class {name}

This command will create a stubbed out class in App\Scrubber\RegexCollection. The Scrubber package will autoload everything from the App\Scrubber\RegexCollection folder with the wildcard value on the regex_loader array in the scrubber config file. You will need to provide a Regex Pattern and a Testable String for the class and you may also provide a Replacement Value if you want to replace the detected value with something other than the default value in the config file.

Validating Regex Patterns

After creating or modifying regex classes, you can verify that every loaded pattern correctly matches its own testable string:

php artisan scrubber:validate

The command tests each pattern against the string returned by getTestableString() and prints a results table:

+---------------------------+--------+---------+
| Pattern                   | Status | Message |
+---------------------------+--------+---------+
| JsonWebToken              | PASS   |         |
| EmailAddress              | PASS   |         |
| SocialSecurityNumber      | PASS   |         |
+---------------------------+--------+---------+
Results: 3 passed, 0 failed, 3 total.

The command exits with a non-zero status when any pattern fails, making it suitable for CI pipelines.

Testing

composer test

yorcreative / laravel-scrubber

Maintainers

Package info

Fund package maintenance!

Statistics

Security