Prompt library

Using my intervals.icu data, summarize my last 7 days of training: total load, time, intensity mix, and the two most notable sessions. Do not invent metrics that aren't available.

Compare my last 14 days of training with the 14 days before that. What changed in load, volume, and intensity distribution? Tell me which icuvisor tool each number came from.

Give me a month-end review for [MONTH]: load trend, polarization of my intensity, longest sessions, and one thing to keep doing and one to change.

Use the `shareable_training_report` prompt to draft a Markdown report for [WEEK/MONTH/RACE PREP/TRAINING JOURNEY] from [START_DATE] to [END_DATE] for [AUDIENCE]. Keep it grounded in my intervals.icu data, cite the icuvisor tool behind key numbers, include a short private-data review checklist, and do not publish, host, upload, or auto-share anything. I will review/redact and share manually if I choose.

How's today looking? Use my intervals.icu data for today only: current CTL/ATL/TSB, wellness, completed activities, planned events with their `workout_status` fields, and any NOTE or race annotations. Keep it concise, call out missing/stale data instead of guessing, and do not describe a planned workout as completed unless icuvisor marks it linked/completed. End with one practical recommendation for the rest of the day.

Look at my wellness for the last 14 days. Summarize sleep, HRV, and resting heart rate trends. Use the correct scales (sleep quality is 1-4) and flag any clear downward trend.

Am I recovered enough for a hard interval session today? Use my recent wellness and my current fitness, fatigue, and form. If today's wellness hasn't synced yet, say so instead of guessing.

Over the last 60 days, did my sleep duration and training load move together? State plainly if the relationship is weak, and do not infer causation.

Find my most recent ride in intervals.icu and explain what happened in it: sport, duration, distance, load, intensity, and any notes. If it was imported from Strava and fields are blank, tell me.

Analyze activity [ACTIVITY]: break down the intervals or laps, time in each zone, and whether the hard efforts held their targets.

For activity [ACTIVITY], tell me which advanced metrics are actually available — decoupling, IF, VI, RPE, feel — and report only those. Do not compute fields the server does not expose.

What are my CTL, ATL, and TSB right now, and how have they moved over the last 6 weeks? Explain what the current form number means for the next few days.

Is my training load trending up, down, or flat over the last 28 days? Say whether the change is meaningful or within normal noise.

Project my fitness if I hold my current weekly load for the next 8 weeks with a recovery week every fourth week. Where do CTL and TSB land?

What are my best efforts and power curve signals over the last 90 days? Compare sprint, threshold, and endurance durations and state the units.

Have my best efforts improved versus the prior 90 days? Break it down by duration bucket and show where I actually gained.

From my ramp test on [DATE], read out my FTP, threshold heart rate, and any VO2max estimate. Use only what's in that activity's data.

Show my planned events for the next 14 days. Resolve the date window in my athlete timezone first with `resolve_calendar_dates`, then flag any week with no easy day and any back-to-back hard sessions.

Compare what I completed this week against what was planned. Use icuvisor's `workout_status` and compliance status counts so pending/future workouts are not treated as misses and skipped/missed caveats stay explicit. Where did I fall short or overshoot, and what is one practical adjustment? Do not change my calendar.

Run a transparent plan-health review for the next 14 days. Resolve the planned window in my athlete timezone with `resolve_calendar_dates`, read my planned events/training plan, compare recent planned-vs-completed adherence using explicit `workout_status` values and compliance status counts, check load/form trajectory and get_fitness_projection assumptions, and caveat missing wellness/readiness data. Treat deload or recovery weeks as intentional unless the evidence says otherwise. If no race event is found, say so and treat any race date I supplied as a scenario anchor. Do not invent a black-box score or write to my calendar; show any proposed change for review first.

What's on my calendar between [DATE] and [DATE]? Group by category and total the planned load.

List the workouts in my intervals.icu library with a one-line summary of each. Group them by folder.

Draft a VO2max bike workout: 5 x 4 min at 110% FTP with 4 min recovery, plus warm-up and cool-down. Show it in intervals.icu workout syntax before saving anything.

Schedule the workout [WORKOUT] from my library onto [DATE]. Confirm the date and load with me before writing.

I have a race on [RACE_DATE]. Review my fitness trend and the planned sessions around race week, then suggest a taper outline. Do not write or delete events.

Give me a post-race retrospective for activity [ACTIVITY]: pacing, decoupling, time in zone, and two lessons for the next race of this type.

For my last 7 rides, compare carbs ingested per hour with each session's load. Fetch them in one get_activities call over the last 30 days, read carbs_ingested_g and load from that payload, and flag underfueled sessions. Where carbs weren't logged, say so — don't estimate.

Review my logged nutrition wellness fields for the last 14 days alongside training load and flag days where intake looked low for the work done.

List the athletes on my roster. For each, give a one-line status: recent load, current form, and any wellness red flag.

Triage athlete [ATHLETE_ID] for this week: recent activities, upcoming planned work, fitness trend, and wellness flags. End with one concise coaching note.

Which of my last 10 activities are missing power, heart rate, or duration data? Give one table row per activity, and identify Strava imports by quoting the source field. If any rows come back truncated, fetch those before answering.

List the advanced icuvisor capabilities available in this session so I know which analyses you can run.