Detecting Suspicious Package Updates
With an increased level of automation provided bypackage managers, which sometimes allow updates to be installedautomatically, malicious package updates are becoming a realthreat in software ecosystems. To address this issue, we proposean approach based on anomaly detection, to identify suspiciousupdates based on security-relevant features that attackers coulduse in an attack. We evaluate our approach in the contextof Node.js/npm ecosystem, to show its feasibility in terms ofreduced review effort and the correct identification of a confirmedmalicious update attack. Although we do not expect it to bea complete solution in isolation, we believe it is an importantsecurity building block for software ecosystems.