Beneath the Mask: Can Contribution Data Unveil Malicious Personas in Open-Source Projects?
On the map: LLM safety and security → Backdoor and data poisoning attacks
Abstract (arXiv)
In February 2024, after building trust over two years with project maintainers by making a significant volume of legitimate contributions, GitHub user "JiaT75" self-merged a version of the XZ Utils project containing a highly sophisticated, well-disguised backdoor targeting sshd processes running on systems with the backdoored package installed. A month later, this package began to be distributed with popular Linux distributions until a Microsoft employee discovered the backdoor while investigating how a recent system upgrade impacted the performance of SSH authentication. Despite its potential global impact, no tooling exists for monitoring and identifying anomalous behavior by personas contributing to other open-source projects. This paper demonstrates how Open Source Intelligence (OSINT) data gathered from GitHub contributions, analyzed using graph databases and graph theory, can efficiently identify anomalous behaviors exhibited by the "JiaT75" persona across other open-source projects.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Backdoor and data poisoning attacks, one of 12 directions in LLM safety and security. Builds on 0 of its 10 references that are in the map. First posted on arXiv 19 Aug 2025.
Builds on
0 of 10 references.
No references here yet.
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2025An LLM-based Quantitative Framework for Evaluating High-Stealthy Backdoor Risks in OSS Supply Chains
- 2025Wolves in the Repository: A Software Engineering Analysis of the XZ Utils Supply Chain Attack
- 2026Not In My Git Yard: Catching Backdoors at Commit and Release Time ↗
- 2026Stylometric Defenses Against Author Impersonation in Software Repositories
- 2023Gotcha! This Model Uses My Code! Evaluating Membership Leakage Risks in Code Models
- 2022Beyond the Surface: Investigating Malicious CVE Proof of Concept Exploits on GitHub
- 2022Identifying Backdoor Attacks in Federated Learning via Anomaly Detection
- 2026LLM-Enabled Open-Source Systems in the Wild: An Empirical Study of Vulnerabilities in GitHub Security Advisories