Best for
- Researchers needing local-first and reproducible bioinformatics workflows
- Users who want an AI agent to orchestrate existing bioinformatics skills
- Developers learning genome-analysis workflows with demo data
The supplied sources identify a PyPI MCP package at version 0.6.1 with a stdio configuration example. The server defaults to demo data and offers an optional local-file access switch; no specific MCP tool names are listed in the supplied manifest.
ClawBio is a bioinformatics-native AI agent skill library covering pharmacogenomics, GWAS, variant interpretation, ancestry, single-cell, and RNA-seq. Its MCP server runs locally over stdio, defaults to demo data, and states that genomic data is not sent off the machine by default.
With Python 3.11+ available, add the configuration to an MCP-compatible client using the provided uvx command. The server uses demo data by default. Set CLAWBIO_MCP_ALLOW_LOCAL_FILES=1 if skills need to read or write local files.
{"mcpServers":{"clawbio":{"command":"uvx","args":["--from","clawbio[mcp]","clawbio","mcp"]}}}