CVE-2026-27959

Red Hat Product Security considers this issue as High severity.

A remote, unauthenticated attacker can send a specially crafted HTTP Host header containing a valid RFC 3986 userinfo component (using the @ delimiter). Due to improper parsing of the authority field in the ctx.hostname API, the application may treat attacker-controlled input as the hostname value. Exploitation occurs when the server processes the malicious request and does not require user interaction (UI:N).

Applications that rely on ctx.hostname to construct absolute URLs, such as: password reset links, email verification links, OAuth redirect URIs, or webhook endpoints, may generate security sensitive URLs that reference an attacker controlled domain. This can result in integrity violations, including manipulation of authentication flows or account takeover scenarios. Integrity impact is therefore rated High.

Confidentiality impact is considered Low because disclosure of sensitive data depends on application-specific usage patterns. The vulnerability does not automatically expose information from the server. However, if an affected application uses the manipulated hostname value to generate security-sensitive links such as password reset or email verification URLs and embeds tokens in those links without additional validation, a victim who later follows such a link may inadvertently disclose those tokens to an attacker controlled domain. Such disclosure is conditional, application-dependent like how the application constructs URLs. Also, the subsequent user interaction beyond the initial exploitation, the confidentiality impact is assessed as Low (C:L).