Code smells are a metaphor to describe issues with source code that are associated with bad program design and bad programming practices which negatively affect software comprehensibility and maintainability. As such, code smells could be ideal indicators to monitor the “health” of a system throughout its evolution, and to guide refactoring efforts to sustain comprehensibility and maintainability. Much code smell related work by both researchers and tool vendors has focused on the formalization and detection of code smells. Recently, studies have started to deeper investigate the impact of smells on software maintenance and evolution. However, there is a long road to go before code smell analysis can cost-efficiently drive exactly those refactorings that will improve the comprehensibility and maintainability of a given system. We identify four challenges to be tackled before we can reach this goal: (1) Lack of a common vocabulary to build a consistent knowledge base, (2) Lack of an ontological framework to compare/synthesize results across studies, (3) Lack of an evaluation framework to compare/evaluate the quality of detection approaches, and (4) Lack of a research agenda to make the knowledge/tools accessible to industrial contexts. The ARCS workshop will bring the members of this community together to establish a common research agenda addressing each of these challenges.