The vulnerability exists in the execute_blender_code function of blender-mcp. An attacker can supply a crafted code argument that is not properly validated, enabling arbitrary code execution via code injection; this inference is based on the description that manipulation of the argument code causes injection. This weakness corresponds to CWE-74 and CWE-94 and can compromise the confidentiality, integrity, and availability of the host system if exploited. The exploit can be performed remotely and has been publicly disclosed, meaning it may be actively used by threat actors.

The product affected is ahujasid’s blender-mcp project. The vulnerability applies to all releases up to the commit 7636d13bded82eca58eb93c3f4cd8708dfdfbe8b and any subsequent unpatched versions, as the project follows a rolling release strategy. No specific version numbers are provided for remediation, so any current instance of the software may be susceptible until a fix is released.

With a CVSS score of 5.1 the vulnerability is considered moderate, but the exploit is possible remotely and publicly available. EPSS information is not available and the vulnerability is not listed in the CISA KEV catalog. Attackers who can chain this injection are able to run arbitrary machine‑level code, potentially escalating privileges or installing persistent backdoors. The risk remains present until an official patch is applied or mitigated by operational controls.