CVE-2026-119052

llama.cpp b2877 before b11935 ggml-rpc Server DoS via Oversized recv_msg Length

Published Oct 29, 2026 CVSS 4.0 · 8.7 High EPSS < 1% KEV · Not listed
Description
llama.cpp b2877 before b11935 contains an uncaught exception vulnerability in the ggml-rpc server that allows unauthenticated remote attackers to crash the rpc-server by sending oversized message lengths. After the HELLO handshake, attackers can send a command such as SET_TENSOR with a length above std::vector::max_size(), triggering an uncaught std::length_error that aborts the process and denies service to all clients.

Analysis

AI-generated

Technical impact

The vulnerability is an uncaught std::length_error triggered by oversized message lengths in the ggml-rpc server. An unauthenticated attacker can send a SET_TENSOR command with a length that exceeds the internal vector maximum, causing the server process to abort. The result is a denial of service to all connected clients, with no exploitation of code execution or data disclosure.

Affected Systems

This issue affects ggml-org's llama.cpp, specifically releases between b2877 and just before b11935. Users running any pre‑b11935 build are vulnerable. The reference commit 7ecb57f7adbea2928aa468ad2aecf05b10725ad4 introduced the problem, and the fix was released in tag b11935.

Risk and Exploitability

The CVSS score of 8.7 indicates a high severity vulnerability. Although EPSS data is not provided, the lack of authentication requirement and remote trigger via the rpc interface suggests that exploitation is feasible from any untrusted network. The vulnerability is not listed in CISA's KEV catalog, but its impact remains significant for services relying on llama.cpp.

Remediation

No solution or workaround provided in the CVE record.

OpenCVE Recommended Actions

  • Upgrade llama.cpp to version b11935 or later to include the patch that prevents oversized recv_msg lengths.
  • If an immediate update is not possible, restrict outbound connections to the ggml-rpc port using firewall rules or network segmentation to limit exposure to trusted hosts.
  • Monitor server logs for repeated aborts and set alerting on process termination to detect exploitation attempts.

Generated by OpenCVE AI on October 29, 2026 at 15:20 UTC.

Tracking

Sign in to view the affected projects.

Advisories

No advisories available
History
Type Values Removed Values Added
Metrics ssvc

{'options': {'Automatable': 'yes', 'Exploitation': 'none', 'Technical Impact': 'partial'}, 'version': '2.0.3'}

Type Values Removed Values Added
First Time appeared Ggml-org
Ggml-org llama.cpp
Vendors & Products Ggml-org
Ggml-org llama.cpp
Type Values Removed Values Added
Description llama.cpp b2877 before b11935 contains an uncaught exception vulnerability in the ggml-rpc server that allows unauthenticated remote attackers to crash the rpc-server by sending oversized message lengths. After the HELLO handshake, attackers can send a command such as SET_TENSOR with a length above std::vector::max_size(), triggering an uncaught std::length_error that aborts the process and denies service to all clients.
Title llama.cpp b2877 before b11935 ggml-rpc Server DoS via Oversized recv_msg Length
First Time appeared Ggml
Ggml llama.cpp
Weaknesses CWE-248
CPEs cpe:2.3:a:ggml:llama.cpp:*:*:*:*:*:*:*:*
Vendors & Products Ggml
Ggml llama.cpp
References
Metrics cvssV3_1

{'score': 7.5, 'vector': 'CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H'}

cvssV4_0

{'score': 8.7, 'vector': 'CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N'}

Weaknesses

Sources

Status
PUBLISHED
Assigner
VulnCheck
Published
Updated
Oct 29, 2026, 5:19 PM
Reserved
Oct 29, 2026, 11:51 AM
Status
Analyzed
Published
Oct 29, 2026, 2:17 PM
Modified
Nov 18, 2026, 6:23 PM
Severity
Important
Public date
Oct 29, 2026, 1:08 PM