Unverified50% confidenceReleaseExact time
gguf_cve_corpus.py (Apache-2.0 license) is responsible for generating test files containing only malformed file headers; it can generate a mini .gguf file containing only the file header for each row in the mapping table, plus a zero-alignment defense-in-depth case.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/15/2026
First Seen
Valid until: 1/13/2027
Sources
Related Entities
Related Claims
Unverifiedgguf_cve_map.json and GGUF_CVE_MAP.md (CC BY 4.0 license) map each GGUF vulnerability advisory to the specific overflowing header field, the CWE number, and the scan rule number that can detect the vulnerability.64% similarVerifiedGGUF is the most mainstream quantization file format in the Ollama and llama.cpp ecosystem, and can be loaded and run directly without additional conversion.62% similarUnverifiedGGUF is a unified file format defined by the llama.cpp project that integrates the previously fragmented GGML format, and Q4_K_M is the de facto standard choice for users with 8 GB of VRAM.61% similarVerifiedllama.cpp was implemented in pure C/C++ by Georgi Gerganov, and the GGUF format packages model architecture metadata, vocabulary, and quantized weights into a single file.61% similarVerifiedThe GGUF format uses a key-value metadata header design with built-in runtime information such as the tokenizer vocabulary, allowing a single file to fully describe a runnable model.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1033714API
curl https://kongchang.com/api/v1/knowledge/claims/1033714MCP
get_claim(id=1033714)