Read the docs
TI-2026-81663 Aliases CVE-2026-118375GHSA-pmf8-m38j-h768

Stack-based buffer overflow in llama.cpp ggml-rpc RDMA transport

Priority MEDIUM 42CWE-787Published · Updated

Key facts

Attack vector
Network
Privileges required
None
User interaction
None
Attack complexity
Low
Attack requirements
Present
Impact type
Code execution
Confidentiality · integrity · availability
High · High · High
Fix status
Fixed

Attack profile from CVSS 4.0 · CNA. Fix status from CNA. Impact type from CWE-787 and the CVSS impact.

Summary

llama.cpp from b8841 before b11865 is affected. An out-of-bounds write in the ggml-rpc RDMA transport function rdma_recv() (CWE-787) lets unauthenticated peers overflow the rpc-server stack. An attacker who advertises RDMA capabilities in the HELLO message can send up to 256 KiB into a 1-byte stack variable, which crashes the server or may allow code execution.

Written with Claude Haiku 5.5 from the sources below · 2026-11-03

Description from the CNA

llama.cpp b8841 before b11865 contains an out-of-bounds write vulnerability in the ggml-rpc RDMA transport rdma_recv() function that allows unauthenticated peers to overflow the rpc-server stack. Attackers advertising RDMA capabilities in the HELLO message can send up to 256 KiB into a 1-byte stack variable in rpc_serve_client(), crashing the server or potentially executing code.

Why it matters now

Exploitation None reported CISA SSVC · not in CISA KEV
EPSS 0.00368 30th percentileScored 2026-11-07
Fix Fixed llama.cpp (ggml-org): b11865

SSVC decision points

CISA ADP · SSVC 2.0.3

Exploitation
None
Automatable
No
Technical impact
Total

The decision (Track, Attend, Act) also depends on your mission and well-being impact; see CISA's SSVC guide.

Affected products

VendorProductAffected versionsFixed inStatusSource
ggml-orgllama.cppb8841 to before b11865b11865Patch availableCNA
  • llama.cpp Vendorggml-org Affected b8841 to before b11865 Fixed in b11865 Patch available CNA

CPE names Threatist (inferred) · 1

Inferred from how NVD named this product before; not given by a source.

CPEVersionsRoleVendor
cpe:2.3:a:ggml:llama.cpp:*:*:*:*:*:*:*:*——ggml-org
  • cpe:2.3:a:ggml:llama.cpp:*:*:*:*:*:*:*:* Versions — — ·ggml-org

Remediation

Update llama.cpp to b11865, the fixed version.

Written with Claude Haiku 5.5 from the sources below · 2026-11-03

  • Fix · CNA llama.cpp (ggml-org) Fixed in b11865

Timeline

  1. 2026-10-28 Threatist record created from CVE-2026-118375
  2. 2026-10-28 Alias added: GHSA-pmf8-m38j-h768
  3. 2026-11-03 CISA SSVC decision points changed

References

Patch 2

Other 3

  • TI-2026-26459 llama.cpp: Unauthenticated RCE via GRAPH_COMPUTE buffer=0 bypass in llama.cpp RPC backend CVSS 3.1CRITICAL 9.8
  • TI-2024-28446 llama.cpp allows Arbitrary Address Read in rpc_server::get_tensor CVSS 3.1CRITICAL 9.8
  • TI-2026-17570 llama.cpp has Out-of-bounds Write in llama-server CVSS 3.1CRITICAL 9.8
  • TI-2024-28447 llama.cpp allows write-what-where in rpc_server::set_tensor CVSS 3.1CRITICAL 10.0
  • TI-2024-15627 Integer overflow in ggml-org llama.cpp CVSS 3.1CRITICAL 9.8
  • TI-2024-14482 Heap-based buffer overflow in ggml-org llama.cpp CVSS 3.1CRITICAL 9.8
See all 52 →

Sources

This product uses the NVD API but is not endorsed or certified by the NVD. Sources and licences