Get the current snapshot of the personal health dashboard (health.ojimpo.com): overall health score/status (NORMAL/CAUTION/CRITICAL) and cultural activity score/status (RICH/MODERATE/LOW), per-category cards (this week vs previous week with change), condition state cards (sleep/readiness/stress/weight/outing/CTL), and trend comments. Scores are percentages of personal baselines (100 = baseline met). Use this first for any 'how am I doing' question.

Get the score time series: one row per date with overall health_score and cultural_score, plus optional per-category score columns. range 1m/3m = daily points, 1y = weekly points. Valid category names — activity: music, exercise, reading, movie, sns, coding, calendar, live, shopping, vitality, outing_activity, cd, podcast, game, like, study; state: sleep, readiness, stress, weight, outing, ctl. Add only the categories you need to keep output small.

List the most recent concrete activities across all sources (e.g. tracks listened, workouts, movies watched, commits) as human-readable lines with relative time. The backend keeps the latest 8.

List all configured data sources with their category, display_type (activity/card_only/state), classification (baseline/event/health_only/both), status, and baseline (base_value per aggregation period). Use to discover valid 'source' and 'category' values for query_records.

Query raw daily activity records (one row per date x source x category: minutes, raw_value, raw_unit). Requires a from/to date range (YYYY-MM-DD). Filter by source and/or category (see list_sources). For ranges over ~2 months prefer group_by week or month to keep output small; raw mode is capped by 'limit' and reports truncation.