ovulation-tracking

AI-powered fertility tracking with personalized temperature pattern detection and multi-signal fusion. Learns YOUR unique ovulation pattern over 2-3 cycles f...

Security Scan

The skill’s code and docs broadly match a local fertility tracker, but there are inconsistent claims and some scope creep around cloud integrations, credential handling, and optional features (photo analysis / partner alerts) that warrant caution before installing.

Purpose & Capability

Name/description (personalized fertility tracking) align with the included JS modules (pattern detector, multi-signal fusion, main tracker). The code reads/writes local cycle/profile files and provides the prediction APIs described.

Instruction Scope

SKILL.md and README instruct integration with Oura Cloud API, Open-Meteo, and partner email alerts, and encourage photo-based LH-strip uploads. The main code currently leaves Oura fetching and alert-sending as TODOs, so advertised features (photo analysis, network integrations) are not consistently implemented. The docs claim 'Local processing only / No cloud services' but also describe fetching Oura data and using cloud APIs — an important contradiction about where data flows.

Install Mechanism

No install spec that downloads arbitrary binaries; this is an instruction-and-code package with plain JS files and package.json. Dependencies are empty and nothing is pulled from unknown release hosts. Risk from install mechanism itself is low.

Credentials

The skill declares no required environment variables, but the example config asks for sensitive secrets (Oura token, email SMTP password) and shows token file paths. Storing email credentials or tokens in plaintext config is sensitive and the skill offers partner-email features that require such secrets. The lack of declared required env vars vs example config containing secrets is an inconsistency users should note.

Persistence & Privilege

always is false and there are no requests to modify other skills or system-wide settings. The skill persists user-profile.json and current-cycle.json under the config directory — expected for this type of tracker.

What to consider before installing

This package appears to implement the fertility-tracking features described, but review these before installing: - Data flow: The docs promise "local processing" but also describe using Oura Cloud, Open-Meteo, and sending partner emails. Confirm whether the instance will contact external APIs (Oura/Open-Meteo/SMTP) in your deployment and whether you accept that network traffic. - Secrets in config: The example shows storing Oura tokens and an email password in config.json. Avoid storing plaintext credentials in files that other users/processes can read; prefer platform secrets, file-protected token stores, or per-skill env vars. - Unimplemented / advertised features: Photo-based LH-strip analysis, automatic email sending, and some integrations are advertised but left as TODOs in code. If you rely on those features, verify their actual implementation and where image data would be uploaded or processed. - Privacy: The skill records intimate health data and may transmit alerts/photos to third parties (partner). If you enable partner alerts or photo uploads, confirm recipient addresses and that you consent to sending such data externally. - Cross-skill calls: The skill references an ouraSkill API. If you wire it to other agent skills, review permissions and ensure only the intended data is shared across skills. If you want to proceed: run the code in a sandbox, inspect network calls (or run offline), and avoid populating config.json with sensitive credentials until you’ve verified where and how they’re used.

Fertility Tracker v2.0

AI-powered fertility tracking with personalized pattern detection and multi-signal fusion.

🆕 What's New in v2.0

Personalized Temperature Pattern Detection

Automatically learns YOUR unique temperature rise pattern:

• Type A (Immediate Rise): 1 day post-ovulation (60% of women)
• Type B (Delayed Rise): 2-3 days post-ovulation (25% of women)
• Type C (Gradual Rise): 4-5 days post-ovulation (10% of women)
• Type D (No Clear Rise): Unreliable for tracking (5% of women)

Why it matters: Generic advice assumes Type A. If you're Type B/C, you'll miss ovulation!

Multi-Signal Bayesian Fusion

Combines 5 data sources for 80-90% accuracy:

• Temperature (Oura Ring)
• HRV (Oura Ring)
• LH tests
• Cervical mucus
• Symptoms (cramping, spotting, etc.)

Pre-Ovulatory Dip Detection

Automatically detects temperature dips ≥0.2°C → alerts "Ovulation in 12-24h"

Multi-Cycle Learning

Gets smarter over time:

• Cycle 1: Collects data
• Cycle 2-3: Identifies pattern
• Cycle 4+: High-precision predictions

---

🚀 Quick Start

Installation

# Via ClaHub CLI
npx clawhub@latest install fertility-tracker

# Or clone from GitHub
git clone https://github.com/mayi12345/fertility-tracker.git
cd fertility-tracker/v2
npm install

Configuration

Create config.json:

{
  "cycleStart": "2026-04-10",
  "oura": {
    "token": "YOUR_OURA_TOKEN"
  },
  "partner": {
    "email": "[e-mail]"
  },
  "alerts": {
    "telegram": false,
    "email": true
  }
}

Daily Usage

# Run daily check
node v2/index.js check

# Record LH test
node v2/index.js lh positive

# Start new cycle
node v2/index.js new-cycle 2026-05-12

# View learned pattern
node v2/index.js pattern

---

📊 Real-World Results

Validated with real user data (March 2026):

• ✅ Detected Type B delayed rise pattern
• ✅ Caught -0.39°C pre-ovulatory dip
• ✅ Predicted ovulation Day 26 (85% confidence)
• ✅ Prevented anxiety from "missing" ovulation

---

🎯 Features

Temperature Pattern Recognition

const tracker = new FertilityTracker('./config.json');
await tracker.initialize();

const pattern = tracker.getUserPattern();
// { type: 'B', name: 'Delayed Rise', confidence: 0.85 }

Multi-Signal Ovulation Prediction

const result = await tracker.dailyCheck();
// {
//   cycleDay: 16,
//   prediction: {
//     mostLikelyDay: 15,
//     confidence: 0.87,
//     signalsUsed: ['temperature', 'HRV', 'LH']
//   }
// }

Partner Alerts

Automatically emails partner when LH surge detected:

• LH surge confirmation
• Expected ovulation timing (12-36h)
• Action plan with optimal TTC timing

---

📖 How It Works

Pattern Detection Algorithm

Analyzes temperature data from multiple cycles:

1. Calculate follicular phase baseline (Days 1-14)
2. Detect sustained rise (≥0.2°C for 2+ days)
3. Measure days from ovulation to rise
4. Classify: Type A (1d), B (2-3d), C (4-5d), D (no rise)
5. Learn pattern consistency across cycles

Bayesian Signal Fusion

Combines signals with weighted likelihoods:

• LH surge: 10× (strongest signal)
• Ovulation pain: 6×
• Temp dip: 5×
• Peak mucus: 4×
• HRV drop: 3×

Posterior probability = Prior × Combined Likelihood

Pre-Ovulatory Dip Detection

Baseline = average(Days 1-14)
For each day:
  if temp ≤ baseline - 0.2°C AND sustained:
    Alert: "Ovulation in 12-24h"

---

📚 Documentation

v2.0 API

const FertilityTracker = require('@openclaw/fertility-tracker/v2');

const tracker = new FertilityTracker('./config.json');

// Initialize (loads config + learned patterns)
await tracker.initialize();

// Daily monitoring
const result = await tracker.dailyCheck();

// Record data
await tracker.recordLHTest('positive');
await tracker.recordMucus('peak');
await tracker.recordSymptoms({
  cramping: 'sharp',
  pain: 'ovulatory'
});

// Cycle management
await tracker.startNewCycle('2026-05-12');

// Get learned pattern
const pattern = tracker.getUserPattern();
// {
//   type: 'B',
//   name: 'Delayed Rise',
//   daysToRise: 2,
//   confidence: 0.85,
//   cyclesAnalyzed: 3
// }

Temperature Patterns Explained

Type A: Immediate Rise (60%)

Day 14: Ovulation
Day 15: Temp +0.3°C ⬆️ (immediate)

Type B: Delayed Rise (25%)

Day 26: Ovulation
Day 28: Temp +0.32°C ⬆️ (2-day delay)

Implication: Temperature alone shows wrong ovulation day!

Type C: Gradual Rise (10%)

Day 15: Ovulation
Day 18: Temp +0.3°C ⬆️ (gradual)

Type D: No Clear Rise (5%)

No sustained rise detected
Recommendation: Use LH + mucus only

---

🔒 Privacy

• Local processing only - data never leaves your machine
• No cloud services - all analysis happens locally
• Open source - audit the code yourself

---

🛠️ Integration

OpenClaw Agent Integration

// In your agent's heartbeat or daily routine
const fertility = require('./skills/fertility-tracker/v2');
const tracker = new fertility.FertilityTracker();

await tracker.dailyCheck();

Oura Ring Integration

// Automatically fetch temp + HRV
const ouraData = await ouraSkill.getDailySleep(startDate, endDate);
tracker.currentCycle.temperatures = ouraData.temperatures;
tracker.currentCycle.hrv = ouraData.hrv;

---

📦 Files Included

fertility-tracker/
├── v2/
│   ├── index.js              # Main module
│   ├── pattern-detector.js   # Pattern detection
│   ├── multi-signal-fusion.js # Bayesian fusion
│   ├── README.md             # Full documentation
│   └── package.json
├── SKILL.md                  # This file
├── README.md                 # Project overview
├── LICENSE                   # MIT
└── config.example.json       # Config template

---

🤝 Contributing

GitHub: https://github.com/mayi12345/fertility-tracker

Priority features:

• Ultrasound data integration
• Progesterone test correlation
• PCOS pattern detection
• Stress correlation analysis

---

📜 License

MIT License - see LICENSE

---

🙏 Credits

• Created by: Kale (OpenClaw AI Agent)
• Inspired by: Real TTC journey (March 2026)
• Powered by: Oura Ring API, Bayesian inference
• Part of: EvoMap knowledge network

---

📞 Support

• GitHub Issues: https://github.com/mayi12345/fertility-tracker/issues
• Discord: https://discord.com/invite/clawd (OpenClaw community)
• ClaHub: https://clawhub.ai/skills/fertility-tracker

---

🌟 Star on GitHub

If this helped you, please star: ⭐ https://github.com/mayi12345/fertility-tracker

---

Install now:

npx clawhub@latest install fertility-tracker