awesome-genomic-skills

This repository is a curated directory of tools and resources for using AI coding assistants to do genomics and bioinformatics work. It organizes entries into a few categories: skill libraries, MCP servers, benchmarks, and general collections of agent skills.

A skill, as the README defines it, is a Markdown file that teaches an AI agent how to perform a specific task. An agent loads these files into its context and can then follow the documented steps when asked. An MCP server, by contrast, is a running service that gives the agent a live connection to external databases or tools, like querying a genetic variant database or running a protein structure lookup. The distinction matters because skills teach procedure while MCP servers provide data access.

The entries span well-known projects from groups like Google DeepMind, OpenAI, and Anthropic, as well as independent academic labs and open-source teams. The skills listed cover common bioinformatics workflows including variant calling from DNA sequencing data, RNA-seq gene expression analysis, single-cell analysis, protein structure work, drug discovery pipelines, and clinical genomics. Several entries include hundreds of pre-built skills for specific analysis types.

MCP server entries give AI agents direct programmatic access to databases like UniProt, Ensembl, gnomAD, PubMed, and others used routinely in life sciences research. Some of the larger bundles combine both skills and MCP servers in one package.

The list also points to genomics-specific benchmarks for evaluating how well AI agents perform on biological analysis tasks, and to broader skill collections that are not specific to life sciences but may be useful in research contexts.

This is a reference and discovery resource, not a software package itself. No installation is needed to browse it, and individual entries link to their respective repositories.