This advisory covers multiple security vulnerabilities affecting Red Hat Enterprise Linux AI 3.3.1, a platform for developing and running Granite family large language models. Among the issues is CVE-2026-25679, where the Go standard library function net/url.Parse incorrectly validates the host/authority component of URLs, accepting invalid URLs by ignoring garbage before an IP-literal. This flaw can cause denial of service. The advisory includes six CVEs in total but does not provide detailed technical descriptions or fixes for all. Red Hat has classified the severity as high but has not released patches yet. The vulnerabilities affect versions >=3.3.0 <=3.3.1 of Red Hat Enterprise Linux AI. No known exploits are currently active in the wild.

The vulnerabilities can impact the stability and security of Red Hat Enterprise Linux AI 3.3.1, potentially allowing denial of service conditions due to improper input validation in core libraries such as Go's net/url package. The advisory does not specify confidentiality or integrity impacts. No active exploitation has been reported. The overall impact is rated high by Red Hat Product Security.

Red Hat advises applying all previously released errata relevant to the system before updating. Currently, no official patches or fixes have been released for these vulnerabilities. For the Go net/url.Parse issue (CVE-2026-25679), no effective mitigation is available that meets Red Hat's criteria for ease of use and applicability. Users should monitor Red Hat advisories for updates and apply fixes once available. No immediate action beyond applying prior errata is recommended at this time.