Skip to content

teoobarca/HealthOS

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

HealthOS

Your health, your system, your rules.

An AI-native health tracking system built on Claude Code. Works for any health condition: cancer, autoimmune, biohacking, diabetes, chronic pain, or anything else. You talk naturally, AI manages everything.


Philosophy

  • AI manages everything — you talk, AI writes structured data, finds patterns, and keeps docs current
  • Evidence-based — research before intervention, citations required, science not opinions
  • Living documentation — files update as new data comes in, nothing goes stale
  • N=1 methodology — formal experiment design for testing interventions on yourself
  • Plain markdown — everything in .md files, git-friendly, no vendor lock-in

Quick Start

git clone https://github.com/teoobarca/HealthOS.git
cd HealthOS
# Open in Claude Code, then:
/setup

That's it. The /setup skill walks you through creating your profile, medications, domains, and customized daily tracking.


How It Works

You say anything naturally. The AI recognizes your intent, loads relevant context, asks clarifying questions, and writes structured data.

Examples:

You Say AI Does
"I had chemo today, felt nauseous after" Logs symptoms, asks severity, checks medication interactions, writes daily log
"How did I sleep?" Checks wearable data (if configured), shows trends, compares to baseline
"I want to try creatine" Designs N=1 experiment with hypothesis, baseline metrics, timeline
"New lab results" Processes data, updates medical records, flags anomalies, updates domains
"Weekly review" Aggregates the week's data, finds patterns, compares to previous week

What You Get

  • Structured health tracking in plain markdown files
  • AI context persistence across sessions (via files, not memory)
  • Research workflows with citation tracking and personalized queries
  • N=1 experiment methodology — formal design for testing interventions
  • Cross-domain pattern detection — AI connects dots between sleep, symptoms, labs, etc.
  • Customizable skills for your specific workflows

Directory Structure

HealthOS/
├── CLAUDE.md              ← AI brain (how the AI behaves)
├── PLAN.md                ← Your health plan and priorities
├── README.md              ← You are here
│
├── docs/
│   ├── context/           ← Your profile, medications, watchlist, connections
│   ├── research/          ← AI-generated research organized by domain
│   └── plans/             ← Design docs and plans
│
├── domains/               ← Health domains (one folder per area)
│   ├── README.md          ← Overview of all domains
│   └── TEMPLATE.md        ← Copy this to create a new domain
│
├── logs/
│   ├── daily/             ← Daily logs (YAML frontmatter + notes)
│   └── weekly/            ← Weekly reviews and trend analysis
│
├── experiments/           ← N=1 experiments
│   └── TEMPLATE.md        ← Experiment design template
│
├── data/
│   ├── inbox/             ← Drop zone for new files (AI processes them)
│   ├── medical/           ← Lab tests, body comp scans, medical records
│   └── reference/         ← Reference materials
│
└── .claude/skills/        ← AI workflow skills

Skills

Skill What It Does
/setup Interactive onboarding — creates your profile, medications, domains, and daily tracking template
/health-research Deep health research with citations, personalized to your condition
/daily-log Create or update a daily log from natural conversation
/weekly-review Weekly summary with trends, patterns, and recommendations
/monthly-review Monthly deep analysis with long-term trends and goal tracking
/add-lab-results Process new lab tests or medical data, update all relevant files
/add-domain Create a new health domain from template
/experiment Design a rigorous N=1 experiment

Adding Integrations

HealthOS ships with no built-in integrations. You add them based on what you use.

How to add a wearable or service:

  1. Write a CLI tool or script that fetches your data
  2. Add the command to the ⚙️ CUSTOMIZE section in CLAUDE.md
  3. Optionally create a skill in .claude/skills/ for complex workflows

Example — adding Oura Ring:

# In CLAUDE.md → CUSTOMIZE section:
## Wearable Integration
- Oura Ring: `cd tools && uv run oura show` (fetches sleep, HRV, activity)
- Note: Activity data is always 1 day behind

Example Domains by Condition

Cancer patient: Treatment Protocol, Side Effects, Labs & Bloodwork, Nutrition, Mental Health, Clinical Trials

Autoimmune: Symptoms, Biomarkers, Triggers, Medications, Flares, Diet

Biohacker: Sleep, Body Composition, Biomarkers, Nutrition, Movement, Cognitive, Supplements

Chronic pain: Pain Tracking, Medications, Physical Therapy, Sleep, Mental Health

Diabetes: Blood Sugar, Medications, Nutrition, Exercise, Complications


Customization

What Where How
AI behavior rules CLAUDE.md Edit the ⚙️ CUSTOMIZE sections
Your profile docs/context/me.md Run /setup or edit directly
Add a domain domains/ Run /add-domain or copy TEMPLATE.md
Add a skill .claude/skills/ Create a folder with SKILL.md
Daily metrics logs/daily/TEMPLATE.md /setup generates this, or edit manually

Requirements

  • Claude Code access
  • Git
  • Basic command line familiarity

Built with Claude Code. Inspired by the data-driven health optimization approach.

About

AI-native health tracking system. Markdown-first, works for any condition. Powered by Claude Code.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors