CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

rampwise-analyze

Activity analysis and injury-risk watch for the Rampwise coaching team. Use for "analyse my last run/ride/swim/brick/strength session", "how did this week compare with last week" (works whether or not there was an activity today), week-over-week or month-over-month change, pace/speed/efficiency trend, "am I at risk of getting injured", "is something off", overuse warning signs, pain check-ins, and before adding load. Compares the latest session with the last similar one, compares the last 7 days with the 7 days before (and the calendar week so far), and raises evidence-labelled injury-risk flags from load, recovery and the user's own pain reports. Flags are prompts to look, never a diagnosis. Works with rampwise-recovery, rampwise-load-injury, rampwise-aerobic-base, rampwise-adapt and rampwise-report. Handles swims (pace per 100 m) and brick sessions: the run after a bike is shown against your last normal run.

Quality

90%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Low

Low-risk findings worth noting

SKILL.md
Quality
Evals
Security

Paths. <skills> below means the folder that contains the rampwise-* skill folders, i.e. the parent of this skill's own folder. Personal data is never inside it: it lives in $RAMPWISE_DATA (default ~/rampwise-data).

Garmin Analyze (activity analysis and injury-risk watch)

Questions you answer: what changed since last week, how did that session compare with the last similar one, and are there warning signs worth acting on? All numbers come from the user's own Garmin data; nothing is compared with other people.

Run

python -X utf8 "<skills>/rampwise-analyze/scripts/week_compare.py"              # last 7 days vs the 7 before; calendar week so far (from Wednesday)
python -X utf8 "<skills>/rampwise-analyze/scripts/latest_activity.py"           # newest session vs the last similar one (--id N for a specific one)
python -X utf8 "<skills>/rampwise-analyze/scripts/injury_watch.py"              # injury-risk flags (add --json for codes; --no-readiness is faster)

Add --json for machine-readable output and --today YYYY-MM-DD to analyse another day.

  • Week over week is always available: the rolling window needs no activity today. The calendar-week comparison is skipped before Wednesday (a Monday "week so far" says nothing). Per sport: minutes, km, sessions, climb, average heart rate, speed or pace, longest session, hard sessions, plus recovery means (HRV, resting HR, sleep) for the two windows.
  • Latest activity: running is shown as pace (min/km), cycling as km/h (or whatever preferences.units says). Aerobic analysis (efficiency, decoupling) comes from rampwise-aerobic-base; it is only readable for steady, comparable sessions. Strength is compared set by set with the last session that shares exercises.
  • Injury watch combines load rules, recovery, and what the user told you (athlete.py check-in, injuries in athlete.json).

The flags (each carries an evidence label; details in references/evidence-map.md)

FlagTriggerLabel
Single-run spikea run more than 10 % over the longest run of the previous 30 days (red above 100 %)research (Frandsen 2025); colours are practice
Week jumprun + ride minutes up 20 % or more on the week before (small bases ignored)practice with mixed support (Nielsen 2014, Damsted 2019; a 10 % rule did not prevent injuries in an RCT, Buist 2008)
No rest day7 days (amber) or 14 days (red) in a rowpractice
Hard days back to backtwo sessions with Garmin aerobic training effect 4.0 or more on consecutive dayspractice
Recovery lowreadiness AMBER or RED from rampwise-recoveryresearch for the signals (Buchheit 2014, Craven 2022); linking them to injury is a gap
Short sleepthe last 3 nights average under 6 h (the 6 h line is practice)limited research (von Rosen 2017, adolescent athletes only); adults: gap
Efficiency droprun efficiency down 5 % or more on the user's own recent baselinepractice
Active or monitored injuryan injury in athlete.json with status active (red) or monitoring (amber); follow its restrictionthe user's own record
Recent injury historyan injury marked resolved less than 60 days ago (information only)research (Johnston 2018, systematic review of 12 endurance studies: a history of injury, medium effect); the 60-day window is practice
Pain check-inthe user reported pain 3 or more (amber), 6 or more (red), or a change in gait (red)practice, in the spirit of the pain-monitoring model (Silbernagel 2007, one small Achilles trial)

The thresholds are in scripts/analysis.py. Some are conventions, not findings; the output says which.

How to talk about it

  • Lead with the comparison ("run minutes down 45 %, ride up 98 %"), then the flags, then one suggested action per flag. Use "was associated with", never "causes".
  • Ask, do not assume. The script lists questions (new pain? where, how bad 0-10, does it change how you walk or run? anything worse the morning after? new shoes, bike setup, surface, illness, stress?). Record answers with athlete.py check-in --area .. --pain N [--gait-change] or injury-add. Pain that changes gait, keeps rising or lasts more than a few days: stop and recommend a clinician.
  • Watch data cannot see an injury. A GREEN watch means "nothing in the data", not "no injury".
  • Never invent a cause. If the week drop is a holiday or illness, the user will say so; ask.

Output to the team (handoff block)

SPECIALIST: analyze
STATUS: ok | partial
KEY: rolling 7 d vs previous 7 d per sport (minutes, km, speed or pace), latest session vs last similar
INJURY_WATCH: GREEN | AMBER | RED
FLAGS: spike, week_jump, no_rest, back_to_back_hard, recovery_low, short_sleep, efficiency_drop, injury_history, pain_checkin
QUESTIONS: what to ask the user
FOR_COACH: "no added load" | "ok to progress"
FOR_ADAPT: flag ids (rampwise-adapt reads them)

Limits

  • Elapsed time includes pauses for some devices; weekly numbers are approximate. Strength is counted in sessions, not load.
  • Pace and speed depend on route, weather and the easy/hard mix; compare similar sessions and say so.
  • Efficiency needs a heart-rate signal; wrist heart rate can drift. Never rank the user against anyone else.

Papers

The studies behind this skill are in references/papers/ (INDEX.md lists the read depth). Before you quote a sample size, effect or limit, open the file; if it is missing or abstract only and the paper may be open access, run paper_fetch.py "<author year>" from the rampwise skill (see its "Papers" section) and say what you are doing. Paywalled papers stay abstract only: say so and claim no more.

Repository

Azahir21/rampwise

Path

skills/rampwise-analyze/SKILL.md

Last updated
First committed

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.