We use cookies to improve your experience on our site.
AVID-2026-R1521
Description
llama.cpp allows write-what-where in rpc_server::set_tensor (CVE-2024-42479)
Details
llama.cpp provides LLM inference in C/C++. The unsafe data pointer member in the rpc_tensor structure can cause arbitrary address writing. This vulnerability is fixed in b3561.
Reason for inclusion in AVID: CVE-2024-42479 describes a write-what-where vulnerability in llama.cpp (rpc_server::set_tensor) enabling arbitrary memory writes. The issue concerns AI software (LLM inference with llama.cpp), is in a component used to deploy/run general-purpose AI systems, is a security vulnerability (RCE/ tampering potential per CVSS), and the report provides explicit signals (CVE, description, fix commit, references). Therefore it meets all AVID criteria for inclusion in AI supply-chain vulnerability curation.
References
- NVD entry
- https://github.com/ggerganov/llama.cpp/security/advisories/GHSA-wcr5-566p-9cwj
- https://github.com/ggerganov/llama.cpp/commit/b72942fac998672a79a1ae3c03b340f7e629980b
Affected or Relevant Artifacts
- Developer: Meta
- Deployer: ggerganov
- Artifact Details:
| Type | Name |
|---|---|
| System | llama.cpp |
Impact
AVID Taxonomy Categorization
- Risk domains: Security
- SEP subcategories: S0100: Software Vulnerability
- Lifecycle stages: L06: Deployment
CVSS
| Version | 3.1 |
| Vector String | CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H |
| Base Score | 10.0 |
| Base Severity | 🔴 Critical |
| Attack Vector | NETWORK |
| Attack Complexity | 🟢 Low |
| Privileges Required | NONE |
| User Interaction | NONE |
| Scope | CHANGED |
| Confidentiality Impact | 🔴 High |
| Integrity Impact | 🔴 High |
| Availability Impact | 🔴 High |
CWE
| ID | Description |
|---|---|
| CWE-123 | CWE-123: Write-what-where Condition |
Other information
- Report Type: Advisory
- Credits:
- Date Reported: 2024-08-12
- Version: 0.3.3
- AVID Entry