Update to Lion Lift

{
  "description": "Review progress and training trends over a given period (days, weeks, or months) across any combination of swim, bike, run, and lift. Use when the user asks for an analysis, review, or summary of their training over a time range, for one sport or all of them.",
  "included_files": [],
  "name": "lion-lift-analyzer",
  "skill_md_contents": "---\nname: lion-lift-analyzer\ndescription: Review progress and training trends over a given period (days, weeks, or months) across any combination of swim, bike, run, and lift. Use when the user asks for an analysis, review, or summary of their training over a time range, for one sport or all of them.\n---\n\n# Analyze training over a period\n\nAnalyze completed training across the sports and time range the user asks about, and explain what the data actually shows — trends, balance, consistency, standout sessions — not just a restated list of numbers.\n\n## Scope the analysis\n\n- Resolve the exact date range and which sports to include (swim, bike, run, lift, or any combination — default to all four if the user doesn't say). Confirm an ambiguous range (\"last month,\" \"this training block\") against today's date rather than guessing a boundary.\n- Use `list_collections` / `list_fields` to confirm the current schema if anything is uncertain. Use `fetch_records` with `start_date`/`end_date` and `status` `completed` for each relevant sport's own collection (`lifts`, `run`, `bike`, `swim`); widen the range slightly around local-day boundaries since dates are stored in UTC.\n- If a `goals` collection exists, pull relevant weekly targets as context for whether training volume matched intent — don't treat a goal as a requirement the user is being graded against.\n\n## Check for the context that actually changes the conclusion\n\n- If a `dailyLog` collection exists, check relevant sleep and work entries for the same period. If short sleep or a heavy work stretch overlaps a dip in training volume or performance, describe it as a possible contributing factor, not a proven cause.\n- Use only available Lion Lift read tools for athlete context. If age, height, weight, sex, or training experience would materially change an interpretation, ask the user directly; never guess or imply these details were retrieved from a profile tool.\n- Note any other real gap in the data (missing heart rate, a sport with too few sessions to say anything meaningful) rather than filling it in.\n\n## Draw real conclusions\n\n- Look for actual patterns: volume trend up or down, consistency (sessions per week, gaps), whether hard efforts cluster or spread out, balance across the requested sports, and any standout session (a clear best/worst effort, a notable jump or drop).\n- Distinguish a real trend (several sessions moving the same direction) from a single outlier session — don't call one unusually good or bad session a trend.\n- Tie conclusions to specific sessions and dates as evidence, not just an aggregate average.\n\n## Present the analysis\n\n- Lead with the clearest two or three takeaways, then the supporting detail per sport. State the exact date range and which sports were included.\n- Call out explicitly, near the top, any contributing factor found in the sleep/work logs or any missing context (age/height/weight, or missing performance data) that limits how confidently a conclusion can be drawn.\n- This skill only reads and analyzes. Never call a scheduling tool from here.\n"
}