CVE-2026-12720

HIGH

Kirki < 6.0.13 - Unauthenticated PHP Object Injection

Title source: CVE List
Search associated code STIX 2.1 vulnerability STIX 2.1 with PoCs

Record summary

CVE-2026-12720 has a selected CVSS score of 7.5 (high); EIP currently links 1 repository PoC.

Description

The Kirki WordPress plugin before 6.0.13 does not restrict which classes may be instantiated when it deserialises data that unauthenticated users can store, leading to PHP Object Injection that is triggered when an administrator later reviews the stored data. With a suitable gadget chain present on the site (via another installed Kirki WordPress plugin before 6.0.13, , or an outdated WordPress version), this could be leveraged to perform a variety of attacks, such as remote code execution.

Description source: CVE List

Exploitation context

Available material

Repository PoCs
1

CISA SSVC decision

ExploitationNone
AutomatableNo
Technical impactTotal

CISA Coordinator · SSVC 2.0.3 · Evaluated Jul 31, 2026 · Source: CVE List

Affected products and versions

1
ProductSourceVersion range

Kirki

Default status: unaffected

CVE ListBefore 6.0.13affected

Proofs of concept

1

Repository PoCs

GitHubwebshellseo8/CVE-2026-12720-Proof-of-ConceptRepository PoCby webshellseo80Stub2 files

5.0 KiB

GitHub
PoC details
AnalysisSuspicious behaviordeepseek-v4-pro:cloud ·

Technical assessment

The artifact is a Python script and README that describe CVE-2026-12720 but contain only a placeholder payload and no functional exploit or scanner logic. The script prints vulnerability information and a hardcoded illustrative serialized string, then exits. It does not connect to a target, send a payload, or verify vulnerability status.

Backdoor review

Suspicious behavior

The Python script is a harmless simulation that prints a placeholder payload and does not perform any network activity or exploitation. However, the README.md materially misrepresents the script's capabilities, claiming it is a 'functional demonstration script' with 'Automated Verification' and 'Customizable Payloads', which directly contradicts the script's actual behavior. This deception about the artifact's functionality raises concern about the operator's intent.

ClassificationStub
Model confidence95%
AuthenticationNot required
Languagespython
Target softwarekirki wordpress plugin
Attack typesphp object injection
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The script contains no code to connect to a target, send a payload, or check for vulnerability. It only prints a placeholder payload and describes the vulnerability. The README claims 'Automated Verification' and 'Customizable Payloads' but the provided Python file implements none of these features. The artifact is materially incomplete for exploitation or scanning.

CVE-2026-12720.py:1-90README.md:19-21

Requirements

  • User must type 'yes' to proceed to the simulation step.CVE-2026-12720.py:80-81

Observed behavior

  • Prints a disclaimer and vulnerability information.CVE-2026-12720.py:36-58
  • Displays a hardcoded placeholder serialized payload without sending it to any target.CVE-2026-12720.py:34CVE-2026-12720.py:61-72
Safety-review evidence

Behaviors behind the backdoor verdict

Flagged behaviors

Deceptive Documentation
  • The README.md describes the artifact as a 'functional demonstration script' with features like 'Automated Verification' and 'Customizable Payloads', but the provided Python script (CVE-2026-12720.py) contains no such functionality; it only prints a hardcoded placeholder payload and requires no target input.README.md:13README.md:19-21CVE-2026-12720.py:61-72

Observables

Url
https://t.me/LatestExploitsA Telegram channel URL embedded in the script as the 'Source' of vulnerability information. It is displayed to the user but not used for any network communication within the provided code.CVE-2026-12720.py:32
Review boundaries

What the analysis did not establish

  • The README describes features (Automated Verification, Customizable Payloads) not present in the supplied Python file.
  • The Python file contains only a placeholder payload and no network operations.
  • Only the two text files (CVE-2026-12720.py and README.md) were reviewed. No other files, dependencies, or network behavior were analyzed.
  • The review cannot determine if the Telegram channel URL is malicious or if the repository contains hidden branches or tags with different content.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

References

2