Instructions to use igfray/gguf-block-count-heap-oob-write-poc with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use igfray/gguf-block-count-heap-oob-write-poc with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="igfray/gguf-block-count-heap-oob-write-poc",
	filename="poc_gguf_block_count_heap_oob_write.gguf",
)

output = llm(
	"Once upon a time,",
	max_tokens=512,
	echo=True
)
print(output)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use igfray/gguf-block-count-heap-oob-write-poc with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf igfray/gguf-block-count-heap-oob-write-poc
# Run inference directly in the terminal:
llama cli -hf igfray/gguf-block-count-heap-oob-write-poc

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf igfray/gguf-block-count-heap-oob-write-poc
# Run inference directly in the terminal:
llama cli -hf igfray/gguf-block-count-heap-oob-write-poc

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf igfray/gguf-block-count-heap-oob-write-poc
# Run inference directly in the terminal:
./llama-cli -hf igfray/gguf-block-count-heap-oob-write-poc

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf igfray/gguf-block-count-heap-oob-write-poc
# Run inference directly in the terminal:
./build/bin/llama-cli -hf igfray/gguf-block-count-heap-oob-write-poc

Use Docker

docker model run hf.co/igfray/gguf-block-count-heap-oob-write-poc

LM Studio
Jan
Ollama
How to use igfray/gguf-block-count-heap-oob-write-poc with Ollama:
```
ollama run hf.co/igfray/gguf-block-count-heap-oob-write-poc
```

Unsloth Studio

How to use igfray/gguf-block-count-heap-oob-write-poc with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for igfray/gguf-block-count-heap-oob-write-poc to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for igfray/gguf-block-count-heap-oob-write-poc to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for igfray/gguf-block-count-heap-oob-write-poc to start chatting

Atomic Chat new
Docker Model Runner
How to use igfray/gguf-block-count-heap-oob-write-poc with Docker Model Runner:
```
docker model run hf.co/igfray/gguf-block-count-heap-oob-write-poc
```

Lemonade

How to use igfray/gguf-block-count-heap-oob-write-poc with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull igfray/gguf-block-count-heap-oob-write-poc

Run and chat with the model

lemonade run user.gguf-block-count-heap-oob-write-poc-{{QUANT_TAG}}

List all available models

lemonade list

You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

llama.cpp GGUF — heap OOB write via unbounded `block_count` PoC (coordinated disclosure)

This repository contains a security-research proof-of-concept file, not a machine-learning model. poc_gguf_block_count_heap_oob_write.gguf is a crafted GGUF that triggers a heap out-of-bounds write during model load in llama.cpp at src/llama-hparams.cpp:11 (set_swa_pattern). It is published solely so the huntr triage team and the llama.cpp / ggml-org maintainers can reproduce the issue during coordinated disclosure.

Impact: Heap out-of-bounds write on model load (memory corruption / DoS). Written values are constrained to 0/1. CVSS 4.0: 7.0 (High).
Affected: llama.cpp master @ 00fa7cb284cbf133fc426733bd64238a3588a33e and current master (guard still absent as of this writing).
Trigger: load the file with any libllama consumer, e.g. llama-cli -m poc_gguf_block_count_heap_oob_write.gguf -p x -n 1. Build with -DLLAMA_SANITIZE_ADDRESS=ON to see the ASAN report.
Inert otherwise — the payload is plain GGUF metadata/bytes; it executes no code and only affects a process that deliberately parses it with a vulnerable llama.cpp.

Files

File	Purpose
`poc_gguf_block_count_heap_oob_write.gguf`	Crafted GGUF (gemma2 arch, `block_count=100000`) that reaches the `set_swa_pattern` OOB write

Downloads last month: 5

GGUF

Model size

0 params

Architecture

gemma2

Hardware compatibility

We're not able to determine the quantization variants.

View all variants

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

You need to agree to share your contact information to access this model

llama.cpp GGUF — heap OOB write via unbounded block_count PoC (coordinated disclosure)

Files

llama.cpp GGUF — heap OOB write via unbounded `block_count` PoC (coordinated disclosure)