A flaw in the Keycloak Services OIDC token introspection endpoint allows a confidential client configured for signed JWT introspection responses to receive the full set of claims of a token it cannot legitimately introspect. The endpoint correctly marks the token as inactive for that client, but still returns the signed JWT that contains sensitive information. As a result, an attacker who can extract a token intended for a different audience can learn confidential data embedded in that token, constituting an information‑disclosure vulnerability.

This issue affects Red Hat Build of Keycloak, Red Hat Single Sign‑On 7, Red Hat Data Grid 8, and Red Hat JBoss Enterprise Application Platform Expansion Pack. No specific version information is supplied in the CNA data, so any installed instance of these products is potentially vulnerable.

The CVSS score of 6.5 categorises the vulnerability as moderate severity. The EPSS score is < 1% and the issue is not listed in CISA’s KEV catalogue. The likely attack vector involves an attacker obtaining a token for a target audience and then sending it to the introspection endpoint as a confidential client; the endpoint will return the signed JWT payload. This requires only the ability to send a normal introspection request, making the exploit straightforward once the system is reachable. The overall risk is therefore moderate, driven by the lack of protection against unauthorized claim disclosure.