SpamPR-Buster
Fighting Spam Pull Requests (2026-2029)
Currently, we notice significant advances in automated and generative AI-based methods to detect and fix bugs and inconsistencies in large-scale software systems. To show the applicability of these novel methods, researchers utilize open source software projects via e.g., on GitHub.
The typical workflow involves researchers discovering various bugs and submitting GitHub issues and pull requests (PRs) to enable the open-source project developers to address or merge them. Consequently, researchers report the fixed issues and accepted pull requests in their publications. The open-source software community generally welcomes the input from the research community to improve the quality of the software systems. However, there has been a growing number of complaints that the sheer volume of issues and pull requests is slowing down progress in open-source software projects, as developers are spending the majority of their time fixing bugs. Consequently, our proposed SpamPR-Buster project aims to understand the categories of PRs and issues filled by automated and generative AI-based methods. We will create a model that classifies and ranks the PR and issues. Furthermore, for important PRs or issues we will develop a method to automatically improve the quality of the PRs or issue, so developers have increased trust when merging the PRs.
Team:
SpamPR-Buster is funded by the German Research Foundation / Deutsche Forschungsgemeinschaft (DFG).