llama.cpp before b2731 Out-of-Bounds Read via GGUF Tensor Offsets
- Published
- Last modified
- Where the record stands: published, rejected by its publisher, or superseded by another id.What do these mean? →
- published
- Whether CISA lists it in its Known Exploited Vulnerabilities catalog: proof that attackers are using it.What do these mean? →
- Not listed in CISA KEV
- Fix
- Fixed in b2731
- Whether working attack code is public (Exploit-DB, Metasploit, Nuclei or GitHub). Public code makes attacks easier.What do these mean? →
- None on Exploit-DB, Metasploit or GitHub
Summary
llama.cpp, an open-source library for running large language models locally, is affected by this flaw when it loads a specially crafted or truncated GGUF model file. An attacker who supplies such a file can cause the program to crash or read memory outside the model file into the model's weights.
Summarized by anthropic/claude-haiku-5.5. Verify details against the original sources.
Read the original description
llama.cpp before b2731 contains an out-of-bounds read vulnerability in llama_model_loader that allows attackers to read memory outside the mapped model file via crafted GGUF tensor offsets. Attackers can supply a crafted or truncated GGUF model file with arbitrary 64-bit tensor offsets to crash the process or load adjacent process memory as model weights.
Timeline
- CVE published
- First seen by cve.now
Affected versions
- ggml-org llama.cpp
From b1046 · fixed in b2731