nightscout-cgm-skill

🥈Silver

The Nightscout CGM Skill uses GitHub Copilot to facilitate blood glucose analysis for diabetes management. This Python-based tool helps users with real-time insights and data visualization to enhance their health monitoring experience.

4770Updated 1mo ago

Intermediate15 minutes to implementdevelopment

Saves ~180 min per use

Quick InstallView Source

claude install shanselman/nightscout-cgm-skill

Works with:

GitHub Copilot

Overview

About This Skill

How to Use

[{"step":"Export your CGM data in a supported format (CSV from Dexcom Clarity, Nightscout API, or Tidepool export).","action":"Use the Nightscout CGM Skill's data ingestion tool to upload your file or connect your Nightscout API endpoint. For GitHub Copilot users, run the command: `nightscout-cgm-skill --import [FILE_PATH] --patient [NAME]`","tip":"Ensure your data includes timestamps, glucose values, insulin doses, and meal carbs. The tool supports most CGM systems (Dexcom, Libre, Medtronic)."},{"step":"Specify your analysis parameters using the prompt template. Include patient details (age, diabetes type), timeframe, and specific goals (e.g., reducing hypoglycemia).","action":"Paste the customized prompt into your AI tool (Claude/ChatGPT) and run it. Example: 'Analyze the last 30 days of blood glucose data for Jamie Lee (Type 1, 34) focusing on reducing overnight lows.'","tip":"For advanced users, include specific patterns to investigate (e.g., 'examine glucose trends after resistance training')."},{"step":"Review the generated report. Pay attention to the 'Key Findings' section and trend visualizations.","action":"Identify the top 2-3 actionable insights. For example, if the report shows frequent 3 AM lows, note the suggested basal rate adjustment.","tip":"Compare the current report with previous months' reports to track progress toward goals."},{"step":"Implement the recommendations in your diabetes management system (pump settings, meal planning, etc.).","action":"For pump users, adjust basal rates and insulin ratios directly in your pump or CGM app. For MDI users, update your insulin dosing schedule in your diabetes log.","tip":"After implementing changes, re-run the analysis in 7-10 days to assess the impact of your adjustments."},{"step":"Share the report with your healthcare provider during your next appointment.","action":"Export the report as a PDF from the Nightscout CGM Skill dashboard. Include any additional context (e.g., recent lifestyle changes) in the notes section.","tip":"Use the 'Trend Analysis' section to discuss long-term patterns with your endocrinologist."}]

Use Cases

Analyze blood glucose data trends

Visualize glucose levels over time

Integrate with diabetes management systems

Provide insights for dietary adjustments

Setup & Installation

Quick Install

Terminal

claude install shanselman/nightscout-cgm-skill

Alternative Install (Git Clone)

git clone https://github.com/shanselman/nightscout-cgm-skill

Requirements

Claude Code or compatible AI agent
Works with: GitHub Copilot

Quick Start Guide

Install the Skill

Copy the install command above and run it in your terminal.

Open Your AI Agent

Launch Claude Code, Cursor, or your preferred AI coding agent.

Try It Out

Use the prompt template or examples below to test the skill.

Customize

Adapt the skill to your specific use case and workflow.

Usage Examples

Prompt Template

Analyze the latest [NUMBER] days of blood glucose data from my Nightscout CGM for [PATIENT_NAME] (age [AGE], diabetes type [TYPE]). Generate a report highlighting: 1) Time in Range (TIR) percentages and trends, 2) Hypoglycemic events (below 70 mg/dL) with timestamps and duration, 3) Hyperglycemic spikes (above 180 mg/dL) with potential causes, 4) Patterns in insulin dosing vs. glucose levels, and 5) Actionable recommendations for [SPECIFIC_GOAL, e.g., 'reducing overnight lows' or 'improving post-breakfast spikes']. Use the data from [DATA_SOURCE, e.g., 'my Dexcom Clarity export' or 'Nightscout API'] and format the output as a markdown table with embedded charts for trends over time. Include a summary of key findings in bullet points at the top of the report.

Example Output

### Nightscout CGM Analysis Report for Alex Carter (Type 1 Diabetes, Age 28)
*Generated on November 15, 2025*

**Key Findings:**
- **Time in Range (70-180 mg/dL):** 68% (Goal: >70%) - *Slightly below target, declining trend over past 30 days*
- **Hypoglycemic Events (Below 70 mg/dL):** 12 events (avg. duration: 22 minutes) - *Peak risk between 2-4 AM*
- **Hyperglycemic Spikes (Above 180 mg/dL):** 8 episodes (avg. peak: 245 mg/dL) - *Most common 30-60 mins post-meal*
- **Insulin Sensitivity:** Pre-breakfast sensitivity factor improved by 12% vs. last month, but post-dinner sensitivity declined by 8%

**Detailed Analysis:**

| Metric | 7-Day Avg | 30-Day Trend | Target | Status |
|--------|-----------|--------------|--------|--------|
| Time in Range | 65% | ▼ -5% | >70% | ⚠️ Below |
| Glucose Management Indicator (GMI) | 7.2% | ▼ -0.3% | <7% | ✅ Good |
| Average Glucose | 168 mg/dL | ▼ -8 mg/dL | <154 mg/dL | ⚠️ High |
| Coefficient of Variation | 38% | ▼ -2% | <36% | ⚠️ High |

**Hypoglycemia Events (Last 7 Days):**
| Date | Time | Duration | Lowest BG | Suspected Cause |
|------|------|----------|-----------|-----------------|
| Nov 12 | 03:15 AM | 28 min | 58 mg/dL | Missed bedtime snack |
| Nov 13 | 11:42 PM | 15 min | 65 mg/dL | Over-correction for 150 mg/dL |
| Nov 14 | 02:30 AM | 22 min | 62 mg/dL | Basal rate too high |

**Hyperglycemia Spikes (Last 7 Days):**
| Date | Time | Peak BG | Meal | Insulin Dose |
|------|------|---------|------|--------------|
| Nov 11 | 12:30 PM | 250 mg/dL | Large pasta lunch | 6.5u (30% under-dosed) |
| Nov 12 | 07:15 PM | 235 mg/dL | Pizza dinner | 5u (25% under-dosed) |

**Trend Analysis (Last 30 Days):**
- **Morning Patterns:** Gradual rise in fasting glucose (145→158 mg/dL) despite consistent basal dosing
- **Exercise Impact:** Post-run glucose drops more aggressively (avg. 15 mg/dL drop per 30 mins exercise)
- **Weekend vs. Weekday:** 18% more hyperglycemic events on weekends (likely higher carb intake)

**Actionable Recommendations:**
1. **Reduce Overnight Lows:** Adjust basal rate from 0.8u/h to 0.7u/h between 1-4 AM. Consider adding a 10g slow-digesting snack at 9:30 PM.
2. **Improve Post-Meal Control:** Increase insulin-to-carb ratio for lunch by 15% and dinner by 10%. Enable dual-wave bolus for high-fat meals.
3. **Exercise Safety:** Reduce pre-exercise insulin by 20% for runs >30 mins. Monitor glucose during and 2 hours post-exercise.
4. **Weekend Strategy:** Plan higher basal rates (+0.1u/h) for Saturday/Sunday mornings. Track weekend carb intake in a food diary.

**Next Steps:**
- Review basal rate adjustments in CareLink/Pump Settings
- Enable CGM trend arrows in Nightscout for real-time alerts
- Schedule endocrinologist appointment to discuss pattern adjustments

*Data Source: Dexcom Clarity Export (Nov 1-15, 2025) | Analysis performed using Nightscout CGM Skill v2.1.4*