Red Hat Keycloak CVE-2026-9083

Exploitation requires an authenticated session with realm administrator privileges specifically holding the 'manage-realm' role within the target Keycloak realm. The vulnerability is triggered through the key provider component creation API endpoint in the Keycloak admin interface - it is not accessible via unauthenticated or end-user-facing endpoints. The attack is network-accessible (AV:N) with low complexity (AC:L), meaning no special network position or race condition is required once credentials are obtained. Exploitation is limited by: (1) mandatory possession of valid 'manage-realm' role credentials, ruling out unauthenticated or low-privilege attackers; (2) the impact being read-only filesystem probing - no write, delete, or code execution capability is exposed by this flaw alone; and (3) the information returned is constrained to file existence and readability by the Keycloak process user, not direct file content exfiltration. Non-default configurations or specific platform versions are not stated as additional prerequisites in the available description.

The CVSS 3.1 vector CVSS:3.1/AV:N/AC:L/PR:H/UI:N/S:U/C:H/I:N/A:N yields a base score of 4.9, which accurately reflects a meaningful but constrained risk. The high-privilege requirement (PR:H - realm administrator with 'manage-realm' role) is the primary limiting factor: this is not an unauthenticated or low-privilege flaw. In practice, exploitation presupposes a compromised or malicious realm administrator account. Within multi-tenant Keycloak deployments, however, realm admins are distinct from super-admins and may be granted to a broader set of operators, making the realistic attack surface wider than PR:H implies for large deployments. The unchanged scope (S:U) and absence of integrity/availability impact (I:N/A:N) confirm this is a pure information disclosure. No EPSS score was provided in the input data, so exploitation probability cannot be quantified. No KEV listing and no POC code are identified, suggesting no active or publicly demonstrated exploitation at time of analysis. The primary real-world risk is reconnaissance: a malicious realm admin can map the host filesystem to locate configuration files, private keys, or credential stores, staging follow-on privilege escalation outside Keycloak.

A threat actor who has obtained realm administrator credentials (via phishing, credential stuffing, or insider access) authenticates to the Keycloak admin API and issues a key provider creation request with the keystore path set to a target such as '/etc/keycloak/standalone/configuration/standalone.xml' or '/root/.ssh/id_rsa'. The Keycloak process attempts to open the specified file, and the API response - whether an error message or a successful parse attempt - reveals whether the file exists and is readable by the Keycloak service account. The attacker iterates over high-value paths to build a filesystem map, identifying credential stores or private key material to support follow-on attacks. No public POC code was identified at time of analysis, but the attack requires only standard admin API calls.