agent-analytics
от dannyshmueli
Web & Frontend DevelopmentКратко
Simple website analytics your AI agent controls end-to-end.
Описание
Что делает этот навык
Web analytics platform designed for AI agents to query via CLI. Tracks page views, custom events, A/B experiments, funnels, retention cohorts, and traffic heatmaps. Agents can set up tracking, run experiments, and interpret results without a dashboard.
Agents can run the full analytics cycle, from setup to experiment to shipping a winner, without switching to a browser dashboard.
Когда его использовать
- Checking weekly traffic trends across multiple projects
- Running A/B tests on hero headlines to improve conversion
- Identifying where users drop off in a signup funnel
- Finding peak traffic hours before scheduling a deploy
- Tracking which CTAs drive the most signups
Пример сценария
User asks: Which of my two hero headlines is converting better?
- 1
Create an A/B experiment: npx @agent-analytics/cli experiments create my-site --name hero_headline --variants control,new_copy --goal cta_click
- 2
Add declarative variant attribute to the hero element in HTML
- 3
Wait for ~100 exposures per variant, then check results: npx @agent-analytics/cli experiments get exp_abc123
- 4
Interpret Bayesian probability_best, lift, and recommendation fields
- 5
Complete the experiment with the winning variant: npx @agent-analytics/cli experiments complete exp_abc123 --winner new_copy
Variant new_copy wins with 4.1% vs 2.3% cta_click rate (79% probability best). Experiment marked complete and winner deployed.
Требования
- AGENT_ANALYTICS_API_KEY environment variable
- Agent Analytics account (sign up at agentanalytics.sh)
Установка
clawhub install dannyshmueli/agent-analyticsВыполните команду CLI или вставьте промпт в Claude Code, Cursor, Codex — любой кодинг-агент, которым пользуетесь.