Welcome to CMIND

A bug localizer for memory bugs in C

CMind stands for C Memory Bug Inspection based on Neural Deduction and is developed by apcl-research at University of Notre Dame. It is specifically developed to localize memory bugs in C. CMind is mainly tested on the o4-mini-2025-04-16 and gpt-5-mini-2025-08-07 models although you can change it to different models with cheaper/better options. You can refer to Open AI for more information.

The authors of CMind are Chia-Yi Su and Collin McMillan. We are commited to keep CMind open source. If you cannot find anything or have any question, please contact us.

The video demonstration for ICPC 2026 is as follows.

Key features

Human-like bug localizer: CMind is designed based on the human study from Human Attention During Localization of Memory Bugs in C Programs.
Flexibility to submit bug report: You can submit the bug report via our online website or deploy it in your local machines
Easy to retrieve results: CMind will return an ID after you submit the bug report. You can use this ID to submit bug report and source code
Combined LLMs with guidance: CMind combines LLMs with guided decision making to avoid LLMs go off-track

Citation

This paper has been accepted by ICPC Demonstration Track. Please cite the following if you use this in your research

@inproceedings{su2026cmind,
  title     = {CMind: An AI Agent for Localizing C Memory Bugs},
  author    = {Chia-Yi Su and Collin McMillan},
  booktitle = {IEEE/ACM International Conference on Software Engineering Tool Demo (ICPC Tool Demo)},
  year      = {2026}
}

About the project

Contributing

When contributing to this repository, please first discuss the change you wish to make via issue, email, or any other method with the owners of this repository before making a change.