This is the one to install if you only install one of the yorkeccak/scientific-skills family. It collapses four separate workflows — PubMed, preprint servers, ClinicalTrials.gov, FDA drug labels — into a single natural-language query.

What it does

A unified semantic search across the canonical biomedical evidence stack. Useful for “what’s the current state of X” queries that span basic research, clinical trials, and regulatory landscape — the exact triangulation a biotech R&D scientist or industry-track PhD does in their head every day, except now you can offload the search step.

Who it’s for

Biotech / pharma R&D scientists — the one-stop scan that maps to how due diligence actually happens
Industry-track PhDs in computational biology, immunology, oncology, neuroscience
Postdocs doing literature-heavy hypothesis generation
Regulatory-adjacent professionals who need to know both the science and the FDA filings

What to watch for

Valyu API dependency for the semantic layer (third-party service)
FDA label coverage is US-specific — for EMA/PMDA work, you’ll still need separate workflows
Clinical trials data freshness depends on registry update cadence — typical delay is days to weeks for major registries
The skill returns a ranked list, not a synthesis. Use a downstream skill or workflow to actually summarize across sources.

Verdict

The closest thing to a “one query for the whole biomedical evidence stack” available as an open-source Claude Skill right now. Strongest for industry R&D contexts where you’re triangulating science + trials + regulatory in the same question.