strava-sync

Strava Training Log Sync

Overview

This skill downloads training data from Strava using the Strava MCP server and creates weekly markdown summaries in the training-log folder. Each week is stored as a single markdown file with detailed activity summaries, including lap details for runs marked as workouts.

Note: This is a Strava-specific data import tool. The running coach system also supports:

• Other platforms (Garmin Connect, Apple Health, Polar Flow, etc.) - athlete provides data
• Manual training logs - athlete creates markdown files directly
• Conversational input - athlete describes training verbally

This skill uses MCP tools specific to Strava (mcp__strava__*). Similar skills could be created for other platforms using their respective APIs or MCP servers.

Workflow

1. Get Week Number from User

First, ask the user to provide the training week number:

• Prompt: "Which training week would you like to sync? (e.g., 1, 2, 3)"
• Store this number as the week identifier
• The file will be saved as week-{n}.md where {n} is this week number

2. Check Strava Connection

Verify that Strava is connected:

# Check connection status

Use the mcp__strava__check-strava-connection tool to verify the connection. If not connected, use mcp__strava__connect-strava to authenticate.

3. Fetch Recent Activities

Retrieve the last 7 days of activities from Strava:

1. Use mcp__strava__get-recent-activities to fetch recent activities (default 30 activities)
2. Filter activities to only those from the last 7 days
3. Calculate the week date range (e.g., "2024-01-15 to 2024-01-21")

4. Process Each Activity

For each activity in the last week:

1. Extract key information:

   • Activity name
   • Date and time
   • Activity type (Run, Ride, Swim, etc.)
   • Distance (convert to miles or km)
   • Duration (format as HH:MM:SS)
   • Elevation gain
   • Average pace/speed
   • Average heart rate (if available)
   • Description/notes

2. Check if the activity is marked as a workout:

   • Look for the workout_type field in the activity data
   • For runs: workout_type 0 = default run, 1 = race, 2 = long run, 3 = workout