RPM CVE-2026-95519

Requires that a user or automation process an attacker-supplied crafted manifest file with 'rpm -q -p' or an equivalent manifest-processing flow (UI:R); the manifest entries are macro-expanded before being opened, so embedded shell commands run. Exploitation is LIMITED by: local delivery of the malicious file (AV:L, no remote network trigger), the requirement that a victim actually invoke rpm against it, and impact scoped to the privileges of the rpm process - meaning root/system compromise only when rpm is run privileged (e.g., in build or CI automation). No special non-default rpm configuration is stated as required.

CVSS 3.1 base score is 7.8 (HIGH) with vector AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H, so despite the CWE-78 command-injection/RCE outcome this is NOT a remotely-reachable network flaw: AV:L plus UI:R means an attacker must get a victim (user or automation) to process an attacker-supplied manifest locally via 'rpm -q -p' or a similar manifest-processing flow. PR:N is accurate (no prior authentication needed to craft the file) but the practical trigger is social/supply-chain delivery of a malicious manifest, not direct attacker access. There is no EPSS score, no CISA KEV listing, and no public exploit code in the provided data, so there is no evidence of active or automated exploitation - this reads as a real but conditional risk rather than a mass-exploited emergency. Severity is genuine where rpm runs as root or inside CI/build automation (full CIA compromise of that account), but exploitation is gated on the UI:R delivery step. Net: patch-worthy given rpm's ubiquity and the code-execution impact, but not a drop-everything internet-facing threat. No key signals conflict with the description; the main caveat is that the affected-product CPE list only enumerates Red Hat products even though the flaw is in rpm itself, which likely under-represents true exposure across other RPM-based distributions.

An attacker crafts a manifest file embedding shell commands inside constructs that rpm macro-expands, then delivers it (e.g., as a package listing, mirror artifact, or CI input). When a user or build/automation pipeline runs 'rpm -q -p' against the file, the embedded commands execute with the privileges of the rpm process - root-level compromise if invoked by a privileged automation job. No public proof-of-concept exploit is identified in the available data.