Code execution via signed-size buffer overflow
Published Jun 17, 2025 · Updated Jun 18, 2025
Buffer overflow in ggml-org llama.cpp before b5662 allows remote attackers to execute code by supplying a malicious GGUF model. The _try_copy helper in src/vocab.cpp casts an oversized token length from size_t to int32_t, bypasses its bounds check, and passes the original size to memcpy. Exploitation requires an application or service to load the attacker-supplied model; successful corruption can crash the process or redirect execution.
Summary
What happened
Buffer overflow in ggml-org llama.cpp before b5662 allows remote attackers to execute code by supplying a malicious GGUF model. The _try_copy helper in src/vocab.cpp casts an oversized token length from size_t to int32_t, bypasses its bounds check, and passes the original size to memcpy. Exploitation requires an application or service to load the attacker-supplied model; successful corruption can crash the process or redirect execution.
The record
- CVE
- CVE-2025-49847
- Published
- Jun 17, 2025
- Updated
- Jun 18, 2025
- Vendor
- Georgi Gerganov
- Product
- llama.cpp
- Classifications
- CWE-119, CWE-195, T1204.002
- Attack vector
- network
- Privileges
- unauthenticated
Timeline
How it unfolded
- Jun 17, 2025CVE publishedPublication date reported by the CVE source.
- Jun 18, 2025Record updatedLatest update available in the CVE record.
Exploitability
Present is not the same as exploitable
Compare your product and version with the public record. A matching version still requires validation against your environment.
Is a vulnerable build present?
Compare these published version ranges with your installed build and any vendor patches.
- Affected versionversion=< b5662
What conditions does exploitation require?
What is affected?
Published CVSS scores
CVSS describes severity. EPSS estimates exploitation probability.
Attacks
What attackers are doing with it
Daily unique IPs observed by Shadowserver honeypots for known exploited vulnerabilities (KEVs). Missing observations do not establish an absence of attacks.
Weakness, pattern, technique
Public exploit references
No public exploit references are available in this record.
Labels summarize the accepted research assessment. They do not indicate a test against your environment.
Technologies
Your stack
See the directory against your own environment.
Your stack
Check the software in your environment
Book a demo to see how Hinoki identifies affected software and validates exploitability in your environment.
Book a demo