GIMP CVE-2026-6695

The user must open a specially crafted PAA (Paint Shop Pro Array) image file in GIMP - this is the sole triggering condition. No authentication, no network service, and no special GIMP configuration are required beyond the PAA plugin being present (it is included by default in RHEL GIMP packages). Exploitation is limited by two concrete factors: (1) UI:R - a human user or automated process must actively open the malicious file in GIMP, making purely opportunistic network exploitation impossible; (2) AV:L - the attacker must deliver the crafted PAA file to the victim's filesystem, typically via phishing, malicious download, or shared storage. Exploitation is further bounded to GIMP sessions running under the victim's user account; no privilege escalation to root is implied by the available data.

The NVD-assigned CVSS 3.1 score of 5.5 (Medium) reflects AV:L/UI:R/S:U/C:N/I:N/A:H - local vector requiring user interaction with no confidentiality or integrity impact. However, this scoring is internally inconsistent with the description and the RCE/Buffer Overflow tags: a heap-based out-of-bounds write that enables heap metadata corruption routinely yields code execution, which should reflect C:H and I:H. This discrepancy should be verified against the Red Hat advisory at https://access.redhat.com/security/cve/CVE-2026-6695. No EPSS data was provided, no KEV listing exists, and no public proof-of-concept has been identified. The AV:L vector and UI:R requirement substantially limit opportunistic exploitation - an attacker cannot exploit this without the user actively opening a crafted PAA file. The real-world risk is moderate: GIMP users who process untrusted image files (creative professionals, CI pipelines doing image processing) are the realistic exposure surface. The RCE potential elevates urgency above what the CVSS score alone communicates.

An attacker crafts a malicious PAA image file with a deliberately malformed LZSS-compressed data stream that causes decode_lzss() to write beyond the heap buffer when the victim opens the file in GIMP. The resulting heap metadata corruption could be leveraged to redirect execution flow and run arbitrary code with the privileges of the user running GIMP. No proof-of-concept exploit has been publicly identified, but the write-what-where primitive inherent in a heap OOB write is a well-understood exploitation primitive on modern systems.