This document discusses automated analysis of CSS rules to support style maintenance. It presents an approach that uses dynamic crawling and analysis to determine the relationship between CSS selectors and the DOM elements they target. By examining CSS rules and properties against actual DOM states at runtime, it can identify matched and unmatched selectors, effective versus ineffective rules, overridden declarations, and undefined CSS classes. This is implemented in a tool called Cilla, which generates reports on examined CSS resources and analysis results. An evaluation aims to determine Cilla's accuracy at detecting unused CSS code compared to other approaches, and quantify typical percentages of unused CSS in web systems.