Skip to content
← Tools / Skills
Skill · .md packageJackZhong2017/10 trustMetrics-checked
Body Root Cause logo

Body Root Cause

Portable SKILL.md for evidence-grounded supplement decisions: one variable at a time, no hype.

Visit official site →

About

Body Root Cause is a single portable SKILL.md that gives AI agents a disciplined workflow for everyday health and supplement questions. Instead of jumping from symptom to supplement stack, it runs structured intake, scans only decision-relevant body systems, updates evidence qualitatively (strong/moderate/weak, never manufactured percentages), and recommends testing one variable at a time with predefined stop rules. It keeps a longitudinal health record, asks consent before writing, and states its boundaries plainly: no diagnosis, no prescription changes, not for children, and clinical evaluation when things are serious.

The scope is deliberately narrow: generally healthy adults with everyday concerns. It won't replace a clinician, and its Bayesian-flavored reasoning is qualitative rather than statistical. But for a 4-star repo, the engineering discipline here (versioned releases, explicit safety boundaries, English canonical records) is unusually careful.

What it can do

  • Structured health-concern intake and triage
  • Qualitative evidence updating without false precision
  • One-variable supplement trial design with stop rules
  • Longitudinal health record with consent gating
  • Explicit safety boundaries and clinical-escalation rules

Setup

Quick install: npx skills add https://github.com/JackZhong2017/body-root-cause-skill --skill body-root-cause. For Claude Code, copy body-root-cause/ to ~/.claude/skills/body-root-cause/. For Codex, copy to ~/.codex/skills/body-root-cause/. For OpenClaw: openclaw skills install https://github.com/JackZhong2017/body-root-cause-skill.

Prerequisites

An AI agent that reads SKILL.md files (Claude Code, Codex, OpenClaw, or similar).

Use cases

nutritionsupplement-guidancehealth-tracking

Works with

Claude CodeCodexopenclaw

Trust score — 7/10

/10 trustMetrics-checked

Scored 7/10 (metrics tier, metrics-v1). Evidence: 4 GitHub stars, MIT license, single portable SKILL.md; Concrete install: npx skills add URL plus Codex/Claude Code/OpenClaw paths; Explicit safety boundaries (no diagnosis, no prescription changes); Caveat: narrow scope (healthy adults, everyday concerns); qualitative not statistical reasoning.

★ 4 GitHub stars

Sources

Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.