Generates personalised nutrition reports from consumer genetic data files.
npx skills add https://github.com/ClawBio/ClawBio --skill nutrigx
Please help me install this Agent Skill. Skill: nutrigx Source: https://github.com/ClawBio/ClawBio/tree/main/skills/nutrigx Command: npx skills add https://github.com/ClawBio/ClawBio --skill nutrigx
Turning raw consumer genetic data (23andMe, AncestryDNA, VCF) into actionable dietary guidance normally requires specialised nutrigenomics knowledge and manual SNP lookup. This Skill automates the entire pipeline locally: it parses the raw genotype file, interrogates a curated panel of nutritionally-relevant SNPs, computes per-nutrient risk scores, and produces a readable report with dietary recommendations — without sending any genetic data to external servers.
Ask the agent to generate a personalised nutrition report from your 23andMe or VCF file, for example: "Generate my personalised nutrition report from genome.csv and flag any folate pathway risks."
High-intent search queries and answers about installing and using this skill.
Run the nutrigx.py script with your 23andMe raw data file as input, for example: python nutrigx.py --input genome.csv --output results/. The parser auto-detects the 23andMe format, extracts panel SNP genotypes, and writes a Markdown report with risk scores and figures.
The parser supports 23andMe .txt/.csv, AncestryDNA .csv, MyHeritage .csv, and standard VCF files. Format is auto-detected from header lines, or you can force it with the --format flag (23andme, ancestry, myheritage, or vcf).
The curated panel covers 28 SNPs across 12 nutrient domains, including MTHFR (folate), VDR and GC (vitamin D), FADS1/FADS2 and APOE (omega-3 and fat), CYP1A2 (caffeine), MCM6 (lactose), and BCMO1 (vitamin A). Definitions live in data/snp_panel.json with literature PMIDs.
No. All computation runs locally on your machine; no genetic data is transmitted anywhere. Output reports contain only gene names, SNP IDs, and risk categories, never raw genotype-level personal data beyond the called genotypes shown in tables.
A warning appears when fewer than 25% of panel SNPs are found in your input file, which happens with chip-to-chip variation between genotyping arrays. SNPs absent from your chip are flagged as not_tested and excluded from scoring, reducing report reliability.
These scores cover only common variants (MAF > 1%) and are not diagnostic or medical advice. Effect sizes derive mostly from European GWAS cohorts, so estimates may not generalise across ancestries, and rare pathogenic variants require clinical confirmation.
ClawBioGenerate personalised nutrition reports from 23andMe, AncestryDNA, or VCF genetic data.
kk580kkIntegrate glucose and nutrition data to manage blood sugar health.
ankitjh4Convert raw 23andMe genotype data into annotated genomic findings with SQLite databases.
mims-harvardMatch patient phenotypes and genetics to rare diseases with HPO terms.
