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.

⚠️ SECURITY RESEARCH ARTIFACT: DO NOT CONVERT

This repository contains deliberately malicious Darknet model files (.cfg + a minimal .weights) used as proof-of-concept for a coordinated-disclosure vulnerability report filed via huntr.com against the darknet2ncnn converter in Tencent/ncnn.

Do NOT run these files through darknet2ncnn on a machine you care about. Each one crashes the converter (SIGFPE, NULL dereference, uncaught exception, or a multi-terabyte allocation attempt).

What they trigger

All crashes are file-controlled, deterministic, and reproduced under AddressSanitizer against a stock build of tools/darknet/darknet2ncnn.cpp at commit dda2e28. All are denial-of-service class.

File Finding Sink Weakness
d1_groups0.cfg divide-by-zero (SIGFPE) darknet2ncnn.cpp:425 parse_cfg CWE-369
d2_nullderef.cfg NULL pointer dereference load_cfg:231 CWE-476
d3_stoi_outofrange.cfg std::out_of_range abort update_field:170 CWE-248
d3b_stoi_invalid.cfg std::invalid_argument abort update_field:174 CWE-1284
d5_hugealloc.cfg excessive / overflowing allocation read_to:817 (load_weights:858) CWE-789/CWE-190
valid.weights shared minimal valid weights header (so the positional weights arg resolves) n/a

Usage under test: ./d2n_asan <cfg> valid.weights out.param out.bin.

Disclosure status

Filed for coordinated disclosure through huntr. Please do not redistribute or public-post these PoCs until maintainer coordination and the disclosure window have completed. These files exist only to let the maintainers and huntr triagers reproduce and fix the issues.

Contact: huntr handle Ano1X8.

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