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.

Log in or Sign Up to review the conditions and access this model content.

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

VC-002 Submission Bundle

Contents:

  • evil_axes_reduce_sum.onnx: minimal malicious ONNX model
  • repro.py: loads the malicious model, runs check_model(), then calls convert_version()
  • Dockerfile: pinned, ASAN-instrumented Linux repro environment
  • run.sh: container entrypoint that preserves the ASAN failure and preload order
  • upload_hf.py: helper to create the HuggingFace repo, enable gating, grant reviewer access, and upload the submission files
  • SUBMISSION.md: Huntr-ready writeup text

Validation target:

  • Upstream repo: https://github.com/onnx/onnx.git
  • Pinned commit: ff324324bb7ef0c8508b1aa402271febbfe39e4e

Expected result:

  • checker accepted model
  • ASAN reports an out-of-bounds read during convert_version()
  • Stack includes onnx/version_converter/adapters/axes_input_to_attribute.h:46
  • Stack includes onnx/version_converter/convert.cc:104
  • Container exits non-zero due to ASAN abort

Affected sibling adapters with the same root cause:

  • onnx/version_converter/adapters/split_13_12.h:45
  • onnx/version_converter/adapters/reshape_5_4.h:45

Usage:

docker build -t onnx-vc-002-submission .
docker run --rm onnx-vc-002-submission

Optional HuggingFace upload:

export HF_TOKEN=...
export HF_REPO_ID=ret2c/onnx-vc-002-poc
python3 upload_hf.py

This helper will:

  • keep the model repo public by default
  • enable gated/manual access by default
  • grant protectai-bot read access by default
  • upload evil_axes_reduce_sum.onnx, Dockerfile, repro.py, run.sh, README.md, and SUBMISSION.md
Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support