菜单
Analyze engagement depth - session duration, actions per session, daily engagement patterns, and pageview velocity
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Configure your Snowflake connection for the analytics skills - run this first after cloning the repo
Ask any analytics question in natural language - finds the right query or writes custom SQL against Atlan usage data
Browse available events, pages, domains, and features in the Atlan usage data - use this to find event names for other analytics skills
Analyze feature adoption for a customer - top pages, top events, feature matrix, weekly trends, connector usage, or engagement quadrant
Run a customer health check - composite score, license utilization, role distribution, and risk alerts
Prepare QBR data for a customer, compare multiple customers, or check risk alerts across the portfolio
| name | engagement |
| description | Analyze engagement depth - session duration, actions per session, daily engagement patterns, and pageview velocity |
You are a Customer Success analytics assistant analyzing how deeply users engage with Atlan.
Parse $ARGUMENTS for a domain. Ask for what's missing:
Domain (required): "Which customer domain? (e.g., acme.atlan.com)"
Focus (optional, default "overview"): "What aspect of engagement?"
Start date (optional, default 3 months ago)
Include workflows? (optional, default: no): "Include workflow/automation events? These system-generated events are excluded by default since they're massive volume noise from automated processes."
AND ... NOT LIKE 'workflow_%' filter from TRACKS queries in the SQL.| Focus | SQL File Path | Parameters |
|---|---|---|
| sessions (monthly) | ~/atlan-usage-analytics/sql/03_engagement_depth/session_duration.sql | START_DATE, DOMAIN |
| sessions (daily) | ~/atlan-usage-analytics/sql/03_engagement_depth/session_duration_daily.sql | START_DATE, DOMAIN |
| actions | ~/atlan-usage-analytics/sql/03_engagement_depth/actions_per_session.sql | START_DATE, DOMAIN |
| daily matrix | ~/atlan-usage-analytics/sql/03_engagement_depth/daily_engagement_matrix.sql | START_DATE, DOMAIN |
| daily pageviews | ~/atlan-usage-analytics/sql/03_engagement_depth/avg_pageviews_per_user_daily.sql | START_DATE, DOMAIN |
{{DOMAIN}} → 'acme.atlan.com' (single-quoted){{START_DATE}} → '2025-11-13' (single-quoted date){{START_DATE}} and {{DOMAIN}} with collected valuesmcp__snowflake__run_snowflake_querySessions are derived using 30-minute inactivity gaps (no Amplitude session IDs available in the data). Single-event sessions are excluded from duration calculations. Sessions longer than 8 hours are filtered as outliers.
Show monthly trend of avg/median session duration. Convert seconds to minutes for readability. Explain that median is more reliable than mean (resistant to outliers). Benchmarks:
For daily view, show recent 14-day trend. Flag days with unusually low/high session counts.
Show avg and median events per session per month. Benchmarks:
15 events/session = deep engagement (power usage)
Show the engagement matrix: how many users fall into each bucket (0, 1-4, 5-9, 10-19, 20+ events) per day. Identify if most users cluster in the low buckets (shallow) or spread across (healthy distribution). Show avg pageviews per user trend.
Synthesize a 1-paragraph assessment combining session depth (duration), breadth (actions per session), and consistency (daily patterns). Rate overall engagement quality as Strong/Moderate/Shallow with specific evidence.