From 1c1308bee4f0748cb7c0006867a26c35025fc16b Mon Sep 17 00:00:00 2001
From: Falguni Moridhara <falgunivmoridhara@gmail.com>
Date: Fri, 17 Jul 2026 20:19:04 +0530
Subject: [PATCH] feat: detect URLs in PII Redaction Middleware

---
 .../src/Defaults/PIIRedactorDefaults.php      |  1 +
 src/PIIRedactor/src/Enums/EntityTypes.php     |  1 +
 src/PIIRedactor/src/PIIRedactor.php           |  4 ++++
 src/PIIRedactor/tests/PIIRedactorTest.php     | 19 +++++++++++++++++++
 src/Support/config/intercept.php              |  1 +
 5 files changed, 26 insertions(+)

diff --git a/src/PIIRedactor/src/Defaults/PIIRedactorDefaults.php b/src/PIIRedactor/src/Defaults/PIIRedactorDefaults.php
index 00603e8..9a4de7e 100644
--- a/src/PIIRedactor/src/Defaults/PIIRedactorDefaults.php
+++ b/src/PIIRedactor/src/Defaults/PIIRedactorDefaults.php
@@ -23,6 +23,7 @@ public static function values(): array
                 'api_key',
                 'bearer_token',
                 'mac_address',
+                'url',
             ],
             'block_entities' => [
                 'credit_card',
diff --git a/src/PIIRedactor/src/Enums/EntityTypes.php b/src/PIIRedactor/src/Enums/EntityTypes.php
index bd0a10a..df69886 100644
--- a/src/PIIRedactor/src/Enums/EntityTypes.php
+++ b/src/PIIRedactor/src/Enums/EntityTypes.php
@@ -24,4 +24,5 @@ enum EntityTypes: string
     case IP_ADDRESS   = 'ip_address';
     case PHONE        = 'phone';
     case MAC_ADDRESS  = 'mac_address';
+    case URL          = 'url';
 }
diff --git a/src/PIIRedactor/src/PIIRedactor.php b/src/PIIRedactor/src/PIIRedactor.php
index 5d3c669..7300637 100644
--- a/src/PIIRedactor/src/PIIRedactor.php
+++ b/src/PIIRedactor/src/PIIRedactor.php
@@ -373,6 +373,10 @@ protected function defaultDetectors(): array
                 EntityTypes::MAC_ADDRESS->value,
                 '/\b(?:[0-9A-Fa-f]{2}[:-]){5}[0-9A-Fa-f]{2}\b/'
             ),
+            new RegexDetector(
+                EntityTypes::URL->value,
+                '/\b(?:https?:\/\/|www\.)[a-zA-Z0-9+&@#\/%?=~_|!:,.;]*[a-zA-Z0-9+&@#\/%=~_|]|\b(?<![a-zA-Z0-9@])(?:[a-zA-Z0-9-]+\.)+[a-zA-Z]{2,}(?:\/[a-zA-Z0-9+&@#\/%?=~_|!:,.;]*[a-zA-Z0-9+&@#\/%=~_|])?\b/'
+            ),
         ];
     }
 
diff --git a/src/PIIRedactor/tests/PIIRedactorTest.php b/src/PIIRedactor/tests/PIIRedactorTest.php
index fec4e43..8bbe6c1 100644
--- a/src/PIIRedactor/tests/PIIRedactorTest.php
+++ b/src/PIIRedactor/tests/PIIRedactorTest.php
@@ -118,6 +118,25 @@ function (AgentPrompt $prompt) use (&$forwardedPrompt): string {
     expect($forwardedPrompt->prompt)->toBe('The router MAC is [MAC_ADDRESS_1]');
 });
 
+it('redacts URLs by default', function (): void {
+    Log::shouldReceive('warning')->once();
+
+    $redactor = new PIIRedactor;
+
+    $forwardedPrompt = null;
+
+    $redactor->handle(
+        makePIIRedactorAgentPrompt('Look at http://example.com, check the page https://example.com, visit the site at www.example.com, or https://127.0.0.1:8080, or view the example.com/dashboard.'),
+        function (AgentPrompt $prompt) use (&$forwardedPrompt): string {
+            $forwardedPrompt = $prompt;
+
+            return 'continued';
+        },
+    );
+
+    expect($forwardedPrompt->prompt)->toBe('Look at [URL_1], check the page [URL_2], visit the site at [URL_3], or [URL_4], or view the [URL_5].');
+});
+
 it('blocks credit cards by default', function (): void {
     Log::shouldReceive('warning')->once();
 
diff --git a/src/Support/config/intercept.php b/src/Support/config/intercept.php
index 9e4959f..95f3b92 100644
--- a/src/Support/config/intercept.php
+++ b/src/Support/config/intercept.php
@@ -90,6 +90,7 @@
                 'api_key',
                 'bearer_token',
                 'mac_address',
+                'url',
             ],
 
             /**