Error Diagnostics Smart Debug

What it does

The Error Diagnostics Smart Debug skill assists in identifying and resolving root causes of test failures and errors. It leverages AI-powered analysis to examine error messages, stack traces, reproduction steps, and other relevant data. The skill then generates ranked hypotheses for the cause of the issue, recommends debugging strategies, and suggests optimal instrumentation locations like breakpoints or logpoints.

When to use it

• When working on error diagnostics tasks.
• When needing guidance or best practices for debugging workflows.
• For analyzing production issues where observability data (logs, traces, metrics) is available.
• To accelerate debugging by generating ranked hypotheses and suggesting testing approaches.

Key capabilities

• Error Pattern Recognition: Identifies common error patterns in stack traces.
• Component Dependency Analysis: Analyzes dependencies between components to pinpoint potential failure points.
• Hypothesis Generation: Generates 3-5 ranked hypotheses for the root cause of an issue, including probability scores and falsification criteria.
• Debugging Strategy Recommendation: Suggests appropriate debugging strategies (interactive, observability-driven, time-travel, chaos engineering, statistical).
• Intelligent Instrumentation: Recommends optimal locations for breakpoints and logpoints.
• Observability Data Integration: Can integrate with error tracking tools (Sentry, Rollbar, Bugsnag), APM platforms (DataDog, New Relic, Dynatrace), tracing systems (Jaeger, Zipkin, Honeycomb), log aggregation tools (ELK, Splunk, Loki) and session replay tools (LogRocket, FullStory).

Example prompts

• "Analyze this error message and stack trace: [paste error information here]"
• "What are the likely causes of these intermittent failures in production?"
• "Suggest a debugging strategy for an issue affecting user cohort X."

Tips & gotchas

• Provide clear goals, constraints, and required inputs to guide the skill's analysis.
• The skill is specifically designed for error diagnostics; it’s not suitable for unrelated tasks or using tools outside its scope.
• For detailed implementation guidance, refer to the resources/implementation-playbook.md file.