shanselman/nightscout-cgm-skill/nightscout-cgm-skill

Nightscout CGM Analysis Skill

This skill provides tools for fetching and analyzing Continuous Glucose Monitor (CGM) data from Nightscout.

⚠️ Before Making Changes

Always run tests before and after modifying `cgm.py`:

cd <skill-path>
python -m pytest tests/ -q           # Quick check (281+ tests)
python -m pytest tests/ --cov=scripts  # With coverage

Available Commands

Run the cgm.py script from this skill's scripts/ directory:

python <skill-path>/scripts/cgm.py <command> [options]

Where <skill-path> is the location where this skill is installed (e.g., ~/.copilot/skills/nightscout-cgm, .github/skills/nightscout-cgm, or .claude/skills/nightscout-cgm).

Commands

\* Optional pump/treatment commands require matching Nightscout uploads. pump and profile require Loop, OpenAPS, AndroidAPS, or similar devicestatus/profile data. treatments, events, and ages require treatment/event entries. CGM-only users won't see errors—commands return a structured message explaining that the data is not available.

Report Command

Generate a comprehensive, self-contained HTML report with interactive charts:

• --days N - Number of days to include (default: 90)
• --output PATH - Custom output path (default: nightscout_report.html)
• --open - Open report in browser after generating

Auto-Sync: Reports automatically sync from Nightscout if local data is more than 30 minutes old.

Report Features:

• Interactive date controls (7d/14d/30d/90d/6mo/1yr/All + custom date pickers)
• All charts recalculate dynamically in browser
• Sticky section navigation for long reports
• Deterministic executive summary with concise status bullets
• Print / Save PDF action with print-friendly styles
• Time-in-Range pie chart
• Modal Day (24-hour profile with percentile bands and target range lines)
• Daily trends, Day of week comparison
• Glucose histogram, Heatmap with hover tooltips
• Weekly summary with week-over-week TIR delta, best day, and context notes
• Key stats: TIR%, GMI (estimated A1C), CV (variability)
• Goal tracking: configurable TIR, CV, GMI, and average glucose targets
• Insulin Delivery (if using Loop/OpenAPS): TDD breakdown (bolus/basal), stacked bar chart, carb tracking

Goals Command

Configure local report goals stored in config.json:

• goals or goals view - Show active goals
• goals set --tir PCT --cv PCT --gmi PCT --average MGDL - Set one or more goals
• goals clear - Remove custom goals and return to defaults
• Goal directions: TIR is minimum target; CV, GMI, and average glucose are maximum targets

Auto-Refresh Command

Configure stale-data sync before read-only queries. This does not run a persistent daemon:

• auto-refresh or auto-refresh view - Show current refresh-on-query settings
• auto-refresh set --minutes N - Sync before queries when the newest local reading is older than N minutes
• auto-refresh on / auto-refresh off - Enable or disable refresh-on-query
• Use refresh --days N for explicit manual sync

Day Command

View detailed readings for a specific date:

• day <date> - Date can be 'today', 'yesterday', '2026-01-16', or 'Jan 16'
• --hour-start H - Start hour for time window (0-23)
• --hour-end H - End hour for time window (0-23)

Worst Command

Find your worst days ranked by peak glucose:

• --days N - Number of days to search (default: 21)
• --hour-start H - Start hour for time window (0-23)
• --hour-end H - End hour for time window (0-23)
• --limit N - Number of worst days to show (default: 5)

Query Options

The query command supports flexible filtering:

• --days N - Number of days to analyze (default: 90)
• --day NAME - Filter by day of week (e.g., Tuesday, or 0-6 where 0=Monday)
• --hour-start H - Start hour for time window (0-23)
• --hour-end H - End hour for time window (0-23)

Chart Options

The chart command creates terminal visualizations:

• --sparkline - Compact trend line using Unicode blocks (▁▂▃▄▅▆▇█)
• --hours N - Hours of data for sparkline (default: 24)
• --date DATE - Specific date for sparkline (e.g., today, yesterday, 2026-01-16)
• --hour-start H - Start hour for sparkline time window (0-23)
• --hour-end H - End hour for sparkline time window (0-23)
• --heatmap - Weekly grid showing time-in-range by day and hour
• --day NAME - Hourly breakdown for a specific day of week
• --color - Use ANSI colors (for direct terminal, not inside Copilot)

Pump and Treatment Commands (Optional)

These commands use optional Nightscout endpoints. The skill auto-detects capabilities and returns a helpful CGM-only message when the required data is not available.

`pump` - Get current pump/loop status:

• IOB (Insulin on Board) and COB (Carbs on Board)
• Predicted glucose trajectory
• Recommended bolus
• Last enacted action (temp basal or bolus)
• Pump status (manufacturer, model, suspended/bolusing)
• Phone battery level

`treatments [--hours N]` - Get recent treatments:

• Boluses (automatic and manual)
• Temp basals
• Carb entries
• Summary totals (total insulin, total carbs)

`events [--tag TEXT] [--days N] [--window-minutes N]` - Correlate existing Nightscout events:

• Matches text in event type, notes, enteredBy, foodType, or reason
• Compares the 30-minute baseline before the event to the post-event window
• Does not store local annotations or edit Nightscout data
• Useful for questions like "what happens after pizza?" when those events already exist in Nightscout

`ages [--count N]` - Get CAGE/SAGE/IAGE:

• CAGE: time since last Site Change
• SAGE: time since last Sensor Change
• IAGE: time since last Insulin Change
• Average interval in days and observed change count per type

`profile` - Get pump profile settings:

• Basal rates by time of day
• Total daily basal
• ISF (Insulin Sensitivity Factor) by time
• Carb ratios by time
• Target glucose ranges
• Loop settings (max bolus, pre-meal targets, override presets)

Examples

# Get current glucose
python scripts/cgm.py current
 
# Generate interactive HTML report (opens in browser)
python scripts/cgm.py report --days 90 --open
 
# Configure report goals
python scripts/cgm.py goals view
python scripts/cgm.py goals set --tir 75 --cv 34 --gmi 6.8 --average 145
 
# Configure refresh-on-query sync
python scripts/cgm.py auto-refresh view
python scripts/cgm.py auto-refresh set --minutes 15
 
# Analyze last 30 days
python scripts/cgm.py analyze --days 30
 
# Find patterns automatically (best/worst times, problem areas)
python scripts/cgm.py patterns
 
# What happened yesterday during lunch (11am-2pm)?
python scripts/cgm.py day yesterday --hour-start 11 --hour-end 14
 
# Find my worst lunch days in the last 3 weeks
python scripts/cgm.py worst --days 21 --hour-start 11 --hour-end 14
 
# What happens on Tuesdays after lunch?
python scripts/cgm.py query --day Tuesday --hour-start 12 --hour-end 15
 
# How are my overnight numbers on weekends?
python scripts/cgm.py query --day Saturday --hour-start 22 --hour-end 6
python scripts/cgm.py query --day Sunday --hour-start 22 --hour-end 6
 
# Morning analysis across all days
python scripts/cgm.py query --hour-start 6 --hour-end 10
 
# Show sparkline of last 24 hours (compact visual trend)
python scripts/cgm.py chart --sparkline --hours 24
 
# Show sparkline for a specific date and time range (with colors)
python scripts/cgm.py chart --date yesterday --hour-start 11 --hour-end 15 --color
 
# Show heatmap of time-in-range by day/hour
python scripts/cgm.py chart --heatmap
 
# Show hourly breakdown for a specific day
python scripts/cgm.py chart --day Saturday
 
# Generate interactive HTML report (opens in browser)
python scripts/cgm.py report --days 90 --open
 
# Compare two time periods
python scripts/cgm.py compare --period1 "last 7 days" --period2 "previous 7 days"
python scripts/cgm.py compare --period1 "this week" --period2 "last week"
 
# Get trend alerts (recurring patterns)
python scripts/cgm.py alerts --days 30
 
# Refresh data from Nightscout
python scripts/cgm.py refresh
 
# Correlate existing Nightscout event/treatment notes
python scripts/cgm.py events --tag pizza --days 90
 
# Get site/sensor/insulin change ages
python scripts/cgm.py ages --count 100

Example Questions You Can Ask

With the pattern analysis capabilities, you can ask natural questions like:

• "What's my current glucose?"
• "Generate a report of my last 90 days"
• "Compare this week to last week"
• "How am I doing vs last month?"
• "What trend alerts do you see?"
• "Analyze my blood sugar for the last 30 days"
• "What patterns do you see in my data?"
• "What's happening on Tuesdays after lunch?"
• "When are my worst times for blood sugar control?"
• "How are my overnight numbers?"
• "When do I tend to go low?"
• "What day of the week is my best for time-in-range?"
• "Show me my morning patterns"
• "Show me a sparkline of my last 24 hours"
• "Show me a heatmap of my glucose"
• "What does Saturday look like?"
• "What was my worst lunch this week?" (you'll be asked what hours you eat lunch)
• "Show me what happened yesterday during dinner"
• "What were my worst days for breakfast the last two weeks?"

Pump-Related Questions (if using Loop/OpenAPS)

• "What's my current IOB?"
• "How much insulin on board do I have?"
• "Show me my pump status"
• "What's my predicted glucose?"
• "What treatments have I had in the last 6 hours?"
• "How much insulin did I take today?"
• "What are my basal rates?"
• "What's my carb ratio?"
• "What are my ISF settings?"
• "Show me my Loop profile"

Output Interpretation

Time in Range (TIR)

• Very Low (<54 mg/dL): Dangerous hypoglycemia
• Low (54-69 mg/dL): Hypoglycemia
• In Range (70-180 mg/dL): Target range
• High (181-250 mg/dL): Hyperglycemia
• Very High (>250 mg/dL): Significant hyperglycemia

Key Metrics

• GMI: Glucose Management Indicator (estimated A1C from CGM data)
• CV: Coefficient of Variation (<36% indicates stable glucose)
• Hourly Averages: Shows patterns throughout the day

Configuration (Required)

Set the NIGHTSCOUT_URL environment variable to your Nightscout API endpoint:

# Linux/macOS
export NIGHTSCOUT_URL="https://your-nightscout-site.com/api/v1/entries.json"
 
# Windows PowerShell
$env:NIGHTSCOUT_URL = "https://your-nightscout-site.com/api/v1/entries.json"