Focus Areas:

• Platform and architecture detection (MVVM, TCA, Clean Architecture)
• Existing telemetry SDK inventory
• Entry points and key user flows
• Network, persistence, and state management layers

Output:

• Platform summary with language/version
• Architecture pattern identification
• Existing SDK coverage assessment
• Gap analysis with priority ranking

instrumentation-reviewer

Reviews code changes for observability issues before they ship.

Focus Areas:

• Anti-patterns (PII leaks, high cardinality, sync telemetry)
• Missing context (no user ID, no session, no screen)
• Naming consistency
• Vendor best practices

Output:

• Issues by severity with file:line references
• Specific fixes with code examples
• Vendor guideline references

Skills

8 skills that activate automatically based on context:

| Skill | Trigger | |-------|---------| | instrumentation-planning | "What should I measure?" | | crash-instrumentation | "How to capture crashes with context" | | session-replay | "Set up session replay" | | interaction-latency | "Track button response time" | | navigation-latency | "Track screen load time" | | network-tracing | "Trace API requests" | | user-journey-tracking | "Track user funnels" | | symbolication-setup | "Configure dSYM upload" |

When to Use

Use this plugin for:

• Adding observability to a new or existing mobile app
• Setting up crash reporting, performance monitoring, or session replay
• Choosing between vendors (Sentry, Datadog, Embrace, Bitdrift, Firebase, etc.)
• Reviewing instrumentation code for anti-patterns
• Understanding what context to attach to errors

Don't use for:

• Backend/server observability
• Web frontend (different patterns)
• General logging questions unrelated to mobile

Directory Structure

mobile-observability/
├── commands/       # /instrument, /audit
├── agents/         # codebase-analyzer, instrumentation-reviewer
├── skills/         # 8 instrumentation skills
├── hooks/          # Anti-pattern warnings for Swift/Kotlin/TypeScript/Dart
└── references/     # 25+ guides covering methodology, platforms, vendors

Philosophy

1. Start with crashes — Get crash reporting and symbolication right before adding anything else
2. Context over volume — One error with full context beats 100 without
3. User intent + system state — Link what users tried to do with device conditions (memory, network, battery)
4. Avoid anti-patterns early — PII leaks and high cardinality are expensive to fix later

Author

Caleb Davis

License

MIT — see LICENSE