Sensor-in-the-Loop Personalized Health Assistant for OpenClaw

What problem does it solve?

Sensor-in-the-loop health coaching turns raw wearable metrics into a state-aware, personalized plan for sleep, recovery, exercise, travel readiness, and work/meeting performance.

Core Features & Use Cases

• Daily health snapshot + 7-day trends: Loads a user’s real CSV data to summarize activity, sleep, and stress/recovery signals with rolling-window context.
• Question classification and context rewriting: Categorizes the user’s query into health decision types and rewrites it into a state-grounded prompt for downstream response generation.
• Structured OpenClaw payload output: Produces a JSON context with answer focus guidance so the LLM can respond in Chinese with prioritized, actionable recommendations.

Use case example: A user asks before a big meeting, “我明天要开会，今天该怎么调整？”. The skill builds daily and 7-day summaries from jian.csv, classifies the question as work/meeting, rewrites the query with sleep/stress/recovery context, and outputs a payload for an LLM-driven actionable plan.

Quick Start

Ask in Chinese for a state-aware recommendation, for example: 我明天要开会，今天该怎么调整？