# Scientific Agent Skills

Scientific Agent Skills is a library of 166 validated Agent Skills, plus more than a hundred scientific databases, giving any compatible AI agent curated workflows for biology, chemistry, medicine and drug discovery.

It is packaged as a portable Agent Plugin, a manifest file plus the skills folder, so plugin-capable clients can load the whole collection at once, and it works with Cursor, Claude Code, Codex, Google Antigravity and any client supporting the open Agent Skills standard. Covered areas include cancer genomics, individual-level genome queries, regulatory-sequence prediction, pathogen-variant surveillance, pharmacokinetic modeling and dose selection, biomedical and regulatory literature retrieval, drug-target binding, molecular dynamics, RNA velocity, microbiome analysis, and geospatial and time-series methods.

A separate desktop application, K-Dense BYOK, runs these skills locally with a user's own API keys and adds web search, file handling and access to the full set of skills and databases, with an option to scale heavy workloads to cloud compute. The project and its skill catalog are described in an arXiv preprint.

Python, MIT licensed.

Website: https://arxiv.org/abs/2609.00065