product-analytics

Router skill. Resolve topic, then Read EXACTLY ONE playbook from routes:. Never preload the cluster — one topic in, one playbook out. [CONFIG]

When to use

A request touches product/business measurement: defining metrics, instrumenting or auditing events, designing or reading an experiment, building cohorts/funnels, or charting results. Not for app feature code, infra, or generic data-engineering ETL — route those to their own clusters. [INFERENCIA]

Inputs → Outputs

In: topic (required; infer from request, ask only if two routes fit), depth (quick=essentials, deep=full playbook with per-step verification).
Out: the single resolved playbook applied to the user's case, with Alfa-core evidence tags ([DOC] [CONFIG] [CÓDIGO] [INFERENCIA] [SUPUESTO]) on every non-obvious claim. [DOC]

Topic disambiguation (pick the right route) [INFERENCIA]

Event schema/naming → analytics-events; wiring SDK/server emit → metrics-instrumentation.
"Did the change work?" → ab-testing; "where do users drop?" → funnel-analytics.
Retention by signup week → cohort-analysis; live dashboards/streaming → real-time-analytics.
"What do we even measure?" → kpi-framework; rendering a chart/figure → data-visualization.

Spine

Discover → Analyze → Execute → Validate. Gates: constitution v6.0.0, evidence tags, script-first. [CONFIG]

Validation gate (done = all true)

Exactly one playbook Read; topic matches the enum verbatim. [CONFIG]
Every statistical/numeric claim carries a tag; [SUPUESTO] paired with a verification step. [DOC]
A/B claims state sample size + significance, never "green = win" before the experiment powers out. [INFERENCIA]

Assets

Routing guard and gate live in assets/ — assets/routing-checklist.md (resolve one topic) and assets/quality-rubric.json (guardian gate). [CONFIG]

Anti-patterns

Loading multiple playbooks; guessing a topic instead of asking when two fit; reporting metrics without instrumentation provenance; calling a test before it reaches significance; inventing benchmark numbers. [SUPUESTO]