What problem does it solve? Mobile apps run on thousands of fragmented devices under unstable networks, making bugs and performance regressions impossible to reproduce locally. This Skill provides guidelines for instrumenting Android and Flutter code so crashes, errors, user behavior, and latency are visible and diagnosable remotely. ## Core Features & Use Cases - Release-safe logging: Configure Timber on Android or Talker on Flutter so debug logs are stripped in release builds while errors still reach the crash reporter. - Crash diagnostics: Attach Crashlytics custom keys and breadcrumbs to captured exceptions so non-fatals are searchable and reproducible. - Analytics and performance traces: Log structured analytics events with bounded cardinality and measure latency with Firebase Performance custom traces and OkHttp/Dio interceptors. - Use Case: When a user reports a failed payment, you can query Crashlytics for the exception filtered by custom keys like payment provider and retry count, then check analytics funnels and network traces to pinpoint whether the cause was a timeout, API error, or cancellation. ## Quick Start Ask the agent to instrument the checkout flow with Talker logging, Crashlytics custom keys on captured exceptions, and a Firebase Analytics event for purchase retries.