Claude Scientific Skills is an open-source collection of 148 ready-to-use agent skills that transform Claude, Cursor, and Codex into powerful scientific research assistants. Created by K-Dense AI, the project provides pre-built workflows spanning 16 scientific domains including bioinformatics, drug discovery, proteomics, clinical research, healthcare AI, materials science, physics, and engineering. The skills integrate with over 250 scientific databases including ChEMBL, UniProt, COSMIC, PubMed, and NCBI, and leverage 55+ optimized Python packages such as BioPython, RDKit, Scanpy, PyTorch Lightning, and scikit-learn. Key workflows include drug discovery pipelines, single-cell RNA-seq analysis, multi-omics biomarker discovery, virtual screening campaigns, clinical variant interpretation, and systems biology network analysis. Each skill includes comprehensive documentation with examples, use cases, and best practices in its SKILL.md file. With 11,200 stars on GitHub and 327 commits, the project has become a cornerstone for researchers looking to accelerate scientific workflows using AI coding assistants. Licensed under MIT, individual skills may carry their own license terms.