Delx Living Body

If you're running multiple wellness MCPs (WHOOP, Oura, Garmin, Strava, etc.) and tired of orchestrating them separately, this meta-server auto-detects which of 15 connectors you've installed and composes parallel calls into one unified layer. Ask "should I train hard today?" and it spawns the right child servers, merges recovery scores, sleep data, and cycle phase, then returns a synthesized recommendation with a rule-based reasoning trace. No LLM calls, no manual config. The six tools include living_body_ask for natural language queries, living_body_compose_context for normalized delx-wellness-context shapes, and living_body_status to see what's detected. Privacy-conscious: strips upstream secrets, caches locally at ~/.delx-living-body/cache.sqlite, never reads child tokens directly.