openidentityplatform/openam · Security

OpenAM has LDAP Injection via `_queryId` Parameter

GHSA ID
GHSA-2vg8-q4c2-5cw3
CVE ID
CVE-2026-41573
Severity
high
Published
Updated
CVSS v4
8.7 · CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N

Description

OpenAM (Open Identity Platform) is an open-source IAM platform providing SSO, OAuth2, SAML, and OpenID Connect capabilities. The CREST REST API layer exposes user query endpoints under /json/{realm}/users. In IdentityResourceV1.queryCollection(), the HTTP query parameter _queryId is passed to a CrestQuery object with escapeQueryId explicitly set to false, bypassing the escape protection introduced as part of the CVE-2021-29156 fix. The unescaped value flows directly to DJLDAPv3Repo.getFilter() where it is concatenated into an LDAP filter string without sanitization, enabling authenticated attackers to inject arbitrary LDAP metacharacters for user enumeration and blind LDAP injection.

Affected Endpoint

Endpoint Auth Required Injection Parameter
GET /openam/json/{realm}/users?_queryId=<INJECTION> SSO Token _queryId
GET /openam/json/{realm}/groups?_queryId=<INJECTION> SSO Token (TBD) _queryId

Background: CVE-2021-29156

CVE-2021-29156 was a pre-authentication LDAP injection in OpenAM's Webfinger endpoint, where user-supplied input reached DJLDAPv3Repo.getFilter() unescaped. The fix introduced the escapeQueryId flag in CrestQuery (defaulting to true) and added Filter.escapeAssertionValue() in the filter-building path:

Credit

Discovered by JD-Security SHENYI Team

CWEs

References