New User Analysis (新用户分析)

One-line positioning: Lock down the "new user" definition with the user, then deliver the right depth of analysis — from a quick cohort check to a full structural-mutation diagnosis. Data source: ThinkingData via ae-cli (event / path / retention / funnel / sql models; SQL is verified for cross-event time alignment — see references/data-validation.md §4.3.5). Output style: 结论先行，结构化表格，红黄绿灯标识；语言跟随用户提问语言。

When to Activate

Trigger phrases: new user definition / caliber / first-day quality / new user cohort / new user structural mutation / new user retention anomaly / channel quality / activation analysis / LTV prediction / behavior sequence / path analysis / first-day funnel.

Not Applicable (Do NOT Activate)

• Existing-user retention/churn → use a dedicated retention-diagnosis Skill.
• Short-drama retention → use the dedicated short-drama retention Skill.
• Creating or defining a new/old user classification property → use new-old-user-define.
• Single-user behavior inspection or individual churn diagnosis → use single-user-inspector.