Design Privacy-Safe Product Analytics with add-analytics

Where this fits

You're building an AI agent that adds a new user-facing feature—say, a "share session" button. Your agent needs to decide whether to log an event, what parameters to include, and how to keep user data private. Without a structured framework, you'll either log everything (creating data noise) or nothing (missing critical product signals). The add-analytics skill gives your agent a proven checklist to design analytics that answer real product questions while respecting privacy.

Why agents benefit

• Enforces six concrete criteria (decision, meaning, authority, population, privacy, reporting) before an event is added, preventing speculative or misleading instrumentation.
• Protects user privacy by requiring bounded enums/booleans or explicitly approved scalars—no code, content, paths, or entity IDs in event parameters.
• Provides authoritative hook points and layered delivery guidance, avoiding double-counting and ensuring the event fires from the correct code seam.
• Includes warehouse verification for BigQuery/Looker, so the agent can confirm that downstream metrics match the product intent.
• Handles edge cases like GoRouter screen reporting and voice/login exceptions, which are common in modern apps but often overlooked.

Practical scenarios

Adding a new feature event

When your agent introduces a "save to playlist" button, it uses the skill to define the event as a confirmed outcome, identify the authority seam (e.g., successful save response), and set bounded parameters like source (enum) and duration (approved scalar).

Changing analytics identity/privacy

If a privacy policy update requires removing user IDs from events, the skill guides your agent through redefining event parameters as anonymous, aggregated values while preserving the event's decision value.

Instrumenting a business outcome

Your agent needs to track "trial-to-paid conversion." The skill helps it define the population (account-level), the authority (subscription activation webhook), and the curated Looker model that will consume the event, ensuring the metric is trustworthy.

Add it to your agent workflow

Install the skill with:

npx skills add sesori-ai/sesori_apps_monorepo --skill add-analytics

Then, whenever your agent is about to add a user-facing feature, invoke the skill and follow the README's six-point checklist. For example, after adding the feature, your agent can run a verification step to ensure each event meets the privacy and reporting rules before committing code.

Compare before adopting

Before integrating this skill, compare it with alternatives: custom event design in your codebase, other analytics skills on OpenAgentSkill, or manual instrumentation guides. Look at quality signals like GitHub stars (105), maintenance freshness (recent commits), and whether the skill's assumptions match your stack (e.g., GoRouter, BigQuery). Also consider how well it fits your existing agent workflow—does it integrate with your code review or deployment pipeline?

Why it is worth tracking

The add-analytics skill stands out because it bridges product thinking and technical implementation—a rare combination in agent skills. Its active repository and clear documentation indicate ongoing maintenance, and the methodology is adaptable beyond Sesori's app. You should evaluate it when your agent starts handling user-facing features that require analytics, especially if privacy compliance is a concern. For more skills like this, visit the OpenAgentSkill listing for add-analytics. Tracking it now ensures your agent is ready to instrument features responsibly from day one.