Аналитика AI-чатбота с HIPAA-защитой

★ 6.8 · general

chatbot-analytics is a Claude Code skill that implements AI chatbot analytics and conversation monitoring while maintaining HIPAA compliance. It tracks 13 key metrics — total sessions, token usage, response latency, fallback rate, topic categories, crisis triggers, and more — collecting only conversation metadata and never storing message content, specific topics, or exact sentiment. The skill provides TypeScript templates for three core operations: tracking conversation start, updating per-message metrics (input/output tokens, response time, fallback flags), and recording conversation outcomes (completed, abandoned, error, crisis_escalated). Category detection works from AI response metadata flags rather than message content, and SQL aggregations feed completion rate, average duration, and crisis escalation counts into dashboards. Built for teams adding chat performance measurement to healthcare or coaching AI products where privacy constraints are non-negotiable.