error-detective skill

Overview

This skill helps you hunt down errors by searching logs and code for error patterns, stack traces, and anomalies. It correlates events across systems to surface likely root causes and gives concrete remediation and prevention steps. Use it proactively when debugging, triaging incidents, or investigating production regressions.

How this skill works

It parses logs with targeted regex and standard stack-trace parsers to extract error signatures and context. It builds timelines, groups similar traces, and correlates occurrences across services, deployments, and time windows to identify cascading failures and spikes. Outputs include extraction patterns, correlation analysis, hypothesized root causes with supporting evidence, and monitoring queries to detect recurrence.

When to use it

• Investigating a sudden increase in error rates or latency in production
• Triage after a deploy to determine whether new code caused failures
• Hunting intermittent or cross-service failures that lack clear origin
• Building monitoring rules to detect reoccurrence of known errors
• Reviewing logs to identify anti-patterns and improve observability

Best practices

• Start with symptom windows and expand time ranges to find patterns
• Normalize timestamps and trace identifiers before correlating events
• Capture full stack traces and surrounding log context for accuracy
• Correlate errors with recent deployments, config changes, and infra events
• Create and store reusable regex and queries for repeated error types

Example use cases

• Extracting stack traces from mixed-language logs and grouping by root exception
• Detecting a cascading failure where one service’s timeout triggers retries and downstream errors
• Generating Splunk/Elasticsearch queries to alert on a specific error fingerprint
• Hypothesizing a race condition by correlating error spikes with traffic surges
• Providing a short remediation plan with code file locations likely responsible

FAQ