Goal Reached Thanks to every supporter — we hit 100%!

Goal: 1000 CNY · Raised: 1359 CNY

100%
Buy Us a Coffee

CVE-2026-78147— ggml-org llama.cpp ggml-RPC Server ggml-rpc.cpp deserialize_tensor deserialization

Quick assessment

Affected
ggml-org llama.cpp
Exploitation
Public or AI PoC available; prioritize validation
Recommended action
Check the vendor advisory and references for a fixed version. If immediate upgrade is impossible, restrict exposure and increase monitoring.

ggml llama.cpp是ggml组织的一个在本地执行大语言模型推理的引擎。 ggml llama.cpp bec4772f6版本存在安全漏洞，该漏洞源于ggml-RPC Server组件中ggml/src/ggml-rpc/ggml-rpc.cpp文件的deserialize_tensor函数对op/op_params参数操作不当，导致反序列化，攻击者可远程发起攻击。

CVSS 7.3 · High EPSS 0.56% · P45

Affected Version Matrix 1

VendorProduct Version RangeStatus
ggml-org llama.cpp bec4772f6 affected
Get alerts for future matching vulnerabilities Log in to subscribe

I. Basic Information for CVE-2026-78147

Vulnerability Information
Have questions about the vulnerability? See if Shenlong's analysis helps!
View Shenlong Deep Dive ↗

Although we use advanced large model technology, its output may still contain inaccurate or outdated information.Shenlong tries to ensure data accuracy, but please verify and judge based on the actual situation.

Vulnerability Title
ggml-org llama.cpp ggml-RPC Server ggml-rpc.cpp deserialize_tensor deserialization
Source: CVE Program / CVE List V5
Vulnerability Description
A vulnerability was found in ggml-org llama.cpp bec4772f6. The impacted element is the function deserialize_tensor of the file ggml/src/ggml-rpc/ggml-rpc.cpp of the component ggml-RPC Server. Performing a manipulation of the argument op/op_params results in deserialization. The attack may be initiated remotely. This vulnerability is distinct from CVE-2026-34159 (GHSA-j8rj-fmpv-wcxw, PR #20908), which only added a buffer==nullptr rejection in create_node() and does not validate op or op_params. The reported GitHub issue was closed automatically due to inactivity.
Source: CVE Program / CVE List V5
CVSS Information
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:L
Source: CVE Program / CVE List V5
Vulnerability Type
可信数据的反序列化
Source: CVE Program / CVE List V5
Vulnerability Title
ggml llama.cpp 输入验证错误漏洞
Source: CNNVD (China National Vulnerability Database)
Vulnerability Description
ggml llama.cpp是ggml组织的一个在本地执行大语言模型推理的引擎。 ggml llama.cpp bec4772f6版本存在安全漏洞，该漏洞源于ggml-RPC Server组件中ggml/src/ggml-rpc/ggml-rpc.cpp文件的deserialize_tensor函数对op/op_params参数操作不当，导致反序列化，攻击者可远程发起攻击。
Source: CNNVD (China National Vulnerability Database)
CVSS Information
N/A
Source: CNNVD (China National Vulnerability Database)
Vulnerability Type
N/A
Source: CNNVD (China National Vulnerability Database)

Affected Products

Vendor Product Affected Versions CPE Subscribe
ggml-org llama.cpp bec4772f6 cpe:2.3:a:ggml-org:llama.cpp:*:*:*:*:*:*:*:*

II. Public POCs for CVE-2026-78147

# POC Description Source Link Shenlong Link
AI-Generated POC Premium
Qwen3.6-35B-A3B · 9174 chars
Pro+ exclusive includes:
Vulnerability reproduction recording (real sandbox build + trigger, exclusive)
In-depth vulnerability mechanism
Trigger conditions & impact
Full executable POC code
Exploit chain & mitigation
POC zip download
100+ AI POC generations per month

III. Intelligence Information for CVE-2026-78147

请登录查看更多情报信息。

Vendor Advisories for CVE-2026-78147 (2)

Proof of Concept for CVE-2026-78147 (1)

Other References for CVE-2026-78147 (1)

IV. Related Vulnerabilities

Same product: llama.cpp
Same vendor: ggml-org
Same weakness: CWE-502

V. Comments for CVE-2026-78147

No comments yet

Leave a comment