The vulnerability exists in the deserialize_tensor function within ggml/src/ggml-rpc/ggml-rpc.cpp of the ggml-RPC Server in the llama.cpp project by ggml-org. An attacker can remotely manipulate the op/op_params arguments during deserialization, potentially leading to unintended behavior. This vulnerability is separate from CVE-2026-34159, which addressed a different issue related to buffer null checks but did not validate op or op_params. The GitHub issue related to this vulnerability was closed due to inactivity, and no official fix or patch has been documented.

The vulnerability allows remote attackers to manipulate deserialization inputs, which may lead to partial compromise of confidentiality, integrity, or availability as indicated by the CVSS vector (VC:L, VI:L, VA:L). The exact impact details are not fully described, but the medium severity score (6.9) reflects a moderate risk. There are no known exploits in the wild at this time.

Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. Since the related GitHub issue was closed due to inactivity and no patch or official fix is documented, users should monitor the ggml-org llama.cpp project for updates. Until a fix is available, consider restricting access to the ggml-RPC Server component to trusted networks to reduce exposure.