Search
Find a vulnerability
Search criteria
Related vulnerabilities
DRUPAL-CONTRIB-2026-134 (CVE-2026-87936)
Vulnerability from drupal – Published: 2026-09-09 17:14 – Updated: 2026-09-11 12:42 – Source website
VLAI
Details
Update 2026-09-11: Increased risk score to reflect publicly documented methods for developing exploits.
This module integrates amazee.ai's AI services into Drupal, including a Postgres/pgvector vector database backend for use with Search API AI Search.
The module doesn't sufficiently sanitize filter values before using them to build SQL queries in its Postgres/pgvector backend, allowing SQL injection.
This vulnerability is mitigated by the fact that a site must be using the module's Postgres/pgvector vector database backend for a Search API AI Search index, and must expose one of that index's non-string fields as a filter (for example, through a View) that is reachable by the attacker.
Credits
References
| URL | Type | |
|---|---|---|
{
"affected": [
{
"database_specific": {
"affected_versions": "\u003c1.3.7 || \u003e=1.4.0 \u003c1.4.3"
},
"package": {
"ecosystem": "Packagist:https://packages.drupal.org/8",
"name": "drupal/ai_provider_amazeeio"
},
"ranges": [
{
"database_specific": {
"constraint": "\u003c1.3.7"
},
"events": [
{
"introduced": "0"
},
{
"fixed": "1.3.7"
}
],
"type": "ECOSYSTEM"
},
{
"database_specific": {
"constraint": "\u003e=1.4.0 \u003c1.4.3"
},
"events": [
{
"introduced": "1.4.0"
},
{
"fixed": "1.4.3"
}
],
"type": "ECOSYSTEM"
}
],
"severity": []
}
],
"aliases": [
"CVE-2026-87936"
],
"credits": [
{
"contact": [
"https://www.drupal.org/u/matank001"
],
"name": "Matan Kotick (matank001)"
}
],
"details": "**Update 2026-09-11:** Increased risk score to reflect publicly documented methods for developing exploits.\n\nThis module integrates amazee.ai\u0027s AI services into Drupal, including a Postgres/pgvector vector database backend for use with Search API AI Search.\n\nThe module doesn\u0027t sufficiently sanitize filter values before using them to build SQL queries in its Postgres/pgvector backend, allowing SQL injection.\n\nThis vulnerability is mitigated by the fact that a site must be using the module\u0027s Postgres/pgvector vector database backend for a Search API AI Search index, and must expose one of that index\u0027s non-string fields as a filter (for example, through a View) that is reachable by the attacker.",
"id": "DRUPAL-CONTRIB-2026-134",
"modified": "2026-09-11T12:42:14.000Z",
"published": "2026-09-09T17:14:00.000Z",
"references": [
{
"type": "WEB",
"url": "https://www.drupal.org/sa-contrib-2026-134"
}
],
"schema_version": "1.7.0"
}