As baseline for a satisfaction of the FAIR4RS principles, research software must be published with metadata in publication repositories that assign persistent identifiers and make the metadata accessible. Additionally, published software metadata must be correct, and rich enough to further improve findability, accessibility, interoperability and reusability. Metadata curation for software publication not only safeguards the respective metadata quality, but also assesses compliance with relevant policies in the Helmholtz Association, its centers, and beyond. Furthermore, software metadata can cumulatively be enriched with dynamic metadata (e.g., usage, citations, development) and can thus be used for evaluation and academic reporting, e.g., to contribute to software-related indicators currently developed within the Helmholtz Association. While software publication can be automated, metadata curation, publication approval and evaluation processes usually require human involvement and should be supported by user interfaces that build on automation tools. In our project, we create “Software CaRD” (Software Curation and Reporting Dashboard), an application that presents software publication metadata for curation. Preprocessed metadata from automated pipelines are made accessible in a structured graphical view. Issues and conflicts are highlighted to allow for easy resolution. Software CaRD also assesses metadata for compliance with configurable policies. For evaluation and reporting, relevant metadata from applicable sources is tracked and visualized. On this poster, we explain the technical solutions employed by the decentralized curation workflow.