Diagnose and Resolve Code Errors

An error-detective skill for log analysis and pattern recognition across distributed systems. Focus areas: log parsing and error extraction via regex patterns, stack trace analysis across languages, error correlation across distributed systems, common error patterns and anti-patterns, log aggregation queries against Elasticsearch and Splunk, and anomaly detection in log streams. Its working method starts from error symptoms and works backward to cause, looks for patterns across time windows, correlates errors with deployments and other changes, checks for cascading failures, and identifies error-rate changes and spikes.

Use when investigating a production error, an unexplained spike in error rate, or a suspected cascading failure across services, especially when the investigation needs to correlate log evidence with a specific deployment or change. Not for tasks unrelated to error or log investigation, and not a substitute for actually fixing the underlying code once a root cause is identified - this skill produces a diagnosis and prevention plan, not the code change itself.

Input is the observed error symptoms plus access to logs, stack traces, and deployment history. Output is a structured investigation package: regex patterns for extracting the error from logs, a timeline of error occurrences, a correlation analysis between affected services, a root-cause hypothesis backed by evidence, monitoring queries to detect recurrence, and the likely code locations causing the errors, covering both an immediate fix and a longer-term prevention strategy.

References Elasticsearch and Splunk as the log-aggregation platforms its queries are built for.

Engineers and SREs debugging a production incident or an unexplained error pattern who need a systematic, evidence-based root-cause investigation rather than guessing from a single stack trace.