How severe is CVE-2026-12491?

CVE-2026-12491 has a CVSS 3.x base score of 4.8, rated medium severity. It is exploitable over network with high attack complexity, requires no privileges and no user interaction. Impact on confidentiality is none, integrity low, and availability low.

Is CVE-2026-12491 being actively exploited?

It is not currently listed in CISA's KEV catalog. Its EPSS exploit-prediction score is 0% (15th percentile), an estimate of the probability of exploitation in the next 30 days.

How do I fix CVE-2026-12491?

Review the linked vendor and NVD advisories for patched versions and mitigations, then upgrade or apply the recommended workaround.

Does CVE-2026-12491 have an EU (EUVD) identifier?

Yes. CVE-2026-12491 is tracked in the ENISA EU Vulnerability Database (EUVD) as EUVD-2026-37645.

When was CVE-2026-12491 published?

CVE-2026-12491 was published on 2026-06-17 and last updated on 2026-07-07.

CVE-2026-12491

CVE-2026-12491 is a medium-severity vulnerability with a CVSS 3.x base score of 4.8. It is not currently listed as actively exploited by CISA, and its EPSS exploit-prediction score is low. The underlying weakness is classified as CWE-115.

Key facts

Severity: Medium (CVSS 3.x base score 4.8)
EPSS exploit prediction: 0% (15th percentile)
Actively exploited: Not listed in CISA KEV
EU (EUVD) id: EUVD-2026-37645
Weakness: CWE-115
Published: 2026-06-17
Last modified: 2026-07-07

Description

A flaw was found in vLLM, an open-source library for large language model inference. This vulnerability arises from improper handling of image metadata, specifically EXIF orientation and PNG transparency (tRNS) data, during image processing. When images are converted to RGB, transparency information may be implicitly discarded or remapped, leading to unexpected rendering of transparent pixels and distortion of input content. This can result in the model misinterpreting image content, potentially affecting the integrity of processed data.

Frequently asked questions

What is CVE-2026-12491?: A flaw was found in vLLM, an open-source library for large language model inference. This vulnerability arises from improper handling of image metadata, specifically EXIF orientation and PNG transparency (tRNS) data, during image processing. When images are converted to RGB, transparency information may be implicitly discarded or remapped, leading to unexpected rendering of transparent pixels and distortion of input content. This can result in the model misinterpreting image content, potentially affecting the integrity of processed data.
How severe is CVE-2026-12491?: CVE-2026-12491 has a CVSS 3.x base score of 4.8, rated medium severity. It is exploitable over network with high attack complexity, requires no privileges and no user interaction. Impact on confidentiality is none, integrity low, and availability low.
Is CVE-2026-12491 being actively exploited?: It is not currently listed in CISA's KEV catalog. Its EPSS exploit-prediction score is 0% (15th percentile), an estimate of the probability of exploitation in the next 30 days.
How do I fix CVE-2026-12491?: Review the linked vendor and NVD advisories for patched versions and mitigations, then upgrade or apply the recommended workaround.
Does CVE-2026-12491 have an EU (EUVD) identifier?: Yes. CVE-2026-12491 is tracked in the ENISA EU Vulnerability Database (EUVD) as EUVD-2026-37645.
When was CVE-2026-12491 published?: CVE-2026-12491 was published on 2026-06-17 and last updated on 2026-07-07.

References

Other CWE-115 vulnerabilities

CVE-2020-27846 — Critical (CVSS 9.8): A signature verification vulnerability exists in crewjam/saml. This flaw allows an attacker to bypass SAML…
CVE-2020-29511 — Critical (CVSS 9.8): The encoding/xml package in Go (all versions) does not correctly preserve the semantics of element namespace prefixes…
CVE-2020-29510 — Critical (CVSS 9.8): The encoding/xml package in Go versions 1.15 and earlier does not correctly preserve the semantics of directives during…
CVE-2020-29509 — Critical (CVSS 9.8): The encoding/xml package in Go (all versions) does not correctly preserve the semantics of attribute namespace prefixes…
CVE-2023-0880 — High (CVSS 8.3): Misinterpretation of Input in GitHub repository thorsten/phpmyfaq prior to 3.1.11.
CVE-2025-5747 — High (CVSS 8.0): WOLFBOX Level 2 EV Charger MCU Command Parsing Misinterpretation of Input Remote Code Execution Vulnerability. This…

Browse all CWE-115 vulnerabilities →