Not all R packages are clearly “good” or “risky”, most fall somewhere in between. This post introduces a scoring framework to help users assess package quality, based on documentation, code, maintenance, and popularity. We also share key principles to ensure the scores are useful, fair, and adaptable to different contexts.