Project Snitch: A Matrix Bridge Leveraging LLMs for Cross-Chat Detection and Reporting of Code Sharing
Faculty Mentor
Antonio Espinoza
Presentation Type
Oral Presentation
Start Date
May 2025
End Date
May 2025
Location
PUB 319
Primary Discipline of Presentation
Computer Science
Abstract
With recent technological developments of LLMs (Large Language Models) and popular chat platforms, the opportunities and methods for violating academic integrity have expanded, making it easier for students to share code and other materials in ways that inhibit fair evaluation and their own learning experience. This paper presents the process, design, and implementation details of Project Snitch, a cross-chat monitoring system that utilizes a LLM using prompt engineering and automated chat accounts (bots) to detect and report instances of code sharing across chat platforms in real time through API calls. The system integrates with two popular chat platforms, Discord and Matrix, to listen to messages on the former, log them, and send an alert to the latter if the LLM model detects code sharing, serving as an automated tool that gathers data on what code sharing is taking place, how the model itself is interpreting it, and alerting the appropriate channels to address it as needed. Additionally, the paper briefly addresses the ethical considerations in the use and deployment of such a system as they relate to user privacy.
Recommended Citation
Locke, Robert B., "Project Snitch: A Matrix Bridge Leveraging LLMs for Cross-Chat Detection and Reporting of Code Sharing" (2025). 2025 Symposium. 4.
https://dc.ewu.edu/srcw_2025/op_2025/o2_2025/4
Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-No Derivative Works 4.0 International License.
Project Snitch: A Matrix Bridge Leveraging LLMs for Cross-Chat Detection and Reporting of Code Sharing
PUB 319
With recent technological developments of LLMs (Large Language Models) and popular chat platforms, the opportunities and methods for violating academic integrity have expanded, making it easier for students to share code and other materials in ways that inhibit fair evaluation and their own learning experience. This paper presents the process, design, and implementation details of Project Snitch, a cross-chat monitoring system that utilizes a LLM using prompt engineering and automated chat accounts (bots) to detect and report instances of code sharing across chat platforms in real time through API calls. The system integrates with two popular chat platforms, Discord and Matrix, to listen to messages on the former, log them, and send an alert to the latter if the LLM model detects code sharing, serving as an automated tool that gathers data on what code sharing is taking place, how the model itself is interpreting it, and alerting the appropriate channels to address it as needed. Additionally, the paper briefly addresses the ethical considerations in the use and deployment of such a system as they relate to user privacy.
