/review — Post-Workout Analysis

1. Get today's date. Read

   knowledge/workout-types.md

   to understand the purpose and targets of the workout type being reviewed. For long runs, also read

   knowledge/long-runs.md

   . For intervals, also read

   knowledge/aerobic-base.md

   .
2. Fetch data using MCP tools (call them directly, in parallel where possible):
   • Activities endpoint for the last 3 days (to find the most recent)
   • Events endpoint for the last 3 days (to find matching planned workout)
3. Identify the most recent activity and fetch its details from the activity endpoint (with intervals)
4. Display:
   • Workout Summary: Name, type, date
   • Planned vs Actual table:

     Metric	Planned	Actual	Diff
     Distance (mi)
     Duration
     Avg Pace (min/mi)
     Elevation (ft)

   • Heart Rate: Avg, max, time in zones (if available)
   • Cadence: Avg (if available)
   • Intervals/Laps: Key splits if interval data exists
   • Training Load: load/TSS from the activity
5. If the activity warrants deeper analysis (tempo runs, intervals, long runs over 2 hours), use the activity streams endpoint to examine:
   • Pace drift: Compare first-half vs second-half average pace from

     velocity_smooth

   • HR decoupling: Compare pace:HR ratio in first half vs second half (decoupling > 5% suggests aerobic ceiling was reached)
   • Elevation profile: Correlate altitude changes with pace/HR to assess climbing efficiency Only fetch streams when the analysis would add value — not for every easy run.
6. Flag any planned-vs-actual deviations > 10%
7. One-line coaching note on the workout execution, grounded in the training science from the knowledge base (e.g., reference HR decoupling thresholds from

   knowledge/aerobic-base.md

   , or pacing principles from

   knowledge/long-runs.md

   )