What problem does it solve? Finding and citing specific evidence across millions of biomedical papers, FDA/PMDA/EMA regulatory filings, clinical trial registries, and protein databases normally requires juggling many separate tools and APIs. This Skill operates the Paperclip CLI from GXL, which exposes all of these corpora as one read-only virtual filesystem with line-numbered documents, so every claim can be cited to an exact line. ## Core Features & Use Cases - Source-scoped semantic search and corpus-wide grep: Search PMC, arXiv, bioRxiv, medRxiv, FDA, trial registries, and UniProt/PDB/ChEMBL with ranking modes, or run full-text regex across the entire corpus to find every paper mentioning a gene or term. - Map/reduce reading at scale: Run LLM readers over a result set to extract the same fields from many papers, then synthesize with reduce strategies like consensus or themes. - Line-pinned citations and claim verification: Every document is line-numbered for precise citation URLs, and opt-in paper repositories verify claims against full text on commit. - Use Case: A researcher asks for a comparison of lipid nanoparticle delivery efficiency across recent in vivo studies. The agent searches PMC, filters the result set, maps an extraction query over 10 papers, and produces an answer where every number links to the exact line it came from. ## Quick Start Ask the agent to use paperclip to search PMC for papers on your topic and summarize the top results with line-level citations.
